Fix: table tag on chunks. (#12126)

- [x] Bug Fix (non-breaking change which fixes an issue)

.github/workflows/tests.yml

@@ -197,37 +197,38 @@ jobs:
           echo -e "COMPOSE_PROFILES=\${COMPOSE_PROFILES},tei-cpu" >> docker/.env
           echo -e "TEI_MODEL=BAAI/bge-small-en-v1.5" >> docker/.env
           echo -e "RAGFLOW_IMAGE=${RAGFLOW_IMAGE}" >> docker/.env
+          sed -i '1i DOC_ENGINE=infinity' docker/.env
           echo "HOST_ADDRESS=http://host.docker.internal:${SVR_HTTP_PORT}" >> ${GITHUB_ENV}
 
           sudo docker compose -f docker/docker-compose.yml -p ${GITHUB_RUN_ID} up -d
           uv sync --python 3.12 --only-group test --no-default-groups --frozen && uv pip install sdk/python --group test
 
-      - name: Run sdk tests against Elasticsearch
+      - name: Run sdk tests against Infinity
         run: |
           export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
           until sudo docker exec ${RAGFLOW_CONTAINER} curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
             echo "Waiting for service to be available..."
             sleep 5
           done
-          source .venv/bin/activate && pytest -s --tb=short --level=${HTTP_API_TEST_LEVEL} test/testcases/test_sdk_api 2>&1 | tee es_sdk_test.log
+          source .venv/bin/activate && DOC_ENGINE=infinity pytest -x -s --tb=short --level=${HTTP_API_TEST_LEVEL} test/testcases/test_sdk_api 2>&1 | tee infinity_sdk_test.log
 
-      - name: Run frontend api tests against Elasticsearch
+      - name: Run frontend api tests against Infinity
         run: |
           export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
           until sudo docker exec ${RAGFLOW_CONTAINER} curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
             echo "Waiting for service to be available..."
             sleep 5
           done
-          source .venv/bin/activate && pytest -s --tb=short sdk/python/test/test_frontend_api/get_email.py sdk/python/test/test_frontend_api/test_dataset.py 2>&1 | tee es_api_test.log
+          source .venv/bin/activate && DOC_ENGINE=infinity pytest -x -s --tb=short sdk/python/test/test_frontend_api/get_email.py sdk/python/test/test_frontend_api/test_dataset.py 2>&1 | tee infinity_api_test.log
 
-      - name: Run http api tests against Elasticsearch
+      - name: Run http api tests against Infinity
         run: |
           export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
           until sudo docker exec ${RAGFLOW_CONTAINER} curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
             echo "Waiting for service to be available..."
             sleep 5
           done
-          source .venv/bin/activate && pytest -s --tb=short --level=${HTTP_API_TEST_LEVEL} test/testcases/test_http_api 2>&1 | tee es_http_api_test.log
+          source .venv/bin/activate && DOC_ENGINE=infinity pytest -x -s --tb=short --level=${HTTP_API_TEST_LEVEL} test/testcases/test_http_api 2>&1 | tee infinity_http_api_test.log
 
       - name: Stop ragflow:nightly
         if: always()  # always run this step even if previous steps failed
@@ -237,35 +238,35 @@ jobs:
 
       - name: Start ragflow:nightly
         run: |
-          sed -i '1i DOC_ENGINE=infinity' docker/.env
+          sed -i '1i DOC_ENGINE=elasticsearch' docker/.env
           sudo docker compose -f docker/docker-compose.yml -p ${GITHUB_RUN_ID} up -d
 
-      - name: Run sdk tests against Infinity
+      - name: Run sdk tests against Elasticsearch
         run: |
           export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
           until sudo docker exec ${RAGFLOW_CONTAINER} curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
             echo "Waiting for service to be available..."
             sleep 5
           done
-          source .venv/bin/activate && DOC_ENGINE=infinity pytest -s --tb=short --level=${HTTP_API_TEST_LEVEL} test/testcases/test_sdk_api 2>&1 | tee infinity_sdk_test.log
+          source .venv/bin/activate && DOC_ENGINE=elasticsearch pytest -x -s --tb=short --level=${HTTP_API_TEST_LEVEL} test/testcases/test_sdk_api 2>&1 | tee es_sdk_test.log
 
-      - name: Run frontend api tests against Infinity
+      - name: Run frontend api tests against Elasticsearch
         run: |
           export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
           until sudo docker exec ${RAGFLOW_CONTAINER} curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
             echo "Waiting for service to be available..."
             sleep 5
           done
-          source .venv/bin/activate && DOC_ENGINE=infinity pytest -s --tb=short sdk/python/test/test_frontend_api/get_email.py sdk/python/test/test_frontend_api/test_dataset.py 2>&1 | tee infinity_api_test.log
+          source .venv/bin/activate && DOC_ENGINE=elasticsearch pytest -x -s --tb=short sdk/python/test/test_frontend_api/get_email.py sdk/python/test/test_frontend_api/test_dataset.py 2>&1 | tee es_api_test.log
 
-      - name: Run http api tests against Infinity
+      - name: Run http api tests against Elasticsearch
         run: |
           export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
           until sudo docker exec ${RAGFLOW_CONTAINER} curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
             echo "Waiting for service to be available..."
             sleep 5
           done
-          source .venv/bin/activate && DOC_ENGINE=infinity pytest -s --tb=short --level=${HTTP_API_TEST_LEVEL} test/testcases/test_http_api 2>&1 | tee infinity_http_api_test.log
+          source .venv/bin/activate && DOC_ENGINE=elasticsearch pytest -x -s --tb=short --level=${HTTP_API_TEST_LEVEL} test/testcases/test_http_api 2>&1 | tee es_http_api_test.log
 
       - name: Stop ragflow:nightly
         if: always()  # always run this step even if previous steps failed

agent/component/agent_with_tools.py

@@ -86,8 +86,9 @@ class Agent(LLM, ToolBase):
         self.tools = {}
         for idx, cpn in enumerate(self._param.tools):
             cpn = self._load_tool_obj(cpn)
-            name = cpn.get_meta()["function"]["name"]
-            self.tools[f"{name}_{idx}"] = cpn
+            original_name = cpn.get_meta()["function"]["name"]
+            indexed_name = f"{original_name}_{idx}"
+            self.tools[indexed_name] = cpn
 
         self.chat_mdl = LLMBundle(self._canvas.get_tenant_id(), TenantLLMService.llm_id2llm_type(self._param.llm_id), self._param.llm_id,
                                   max_retries=self._param.max_retries,
@@ -95,7 +96,12 @@ class Agent(LLM, ToolBase):
                                   max_rounds=self._param.max_rounds,
                                   verbose_tool_use=True
                                   )
-        self.tool_meta = [v.get_meta() for _,v in self.tools.items()]
+        self.tool_meta = []
+        for indexed_name, tool_obj in self.tools.items():
+            original_meta = tool_obj.get_meta()
+            indexed_meta = deepcopy(original_meta)
+            indexed_meta["function"]["name"] = indexed_name
+            self.tool_meta.append(indexed_meta)
 
         for mcp in self._param.mcp:
             _, mcp_server = MCPServerService.get_by_id(mcp["mcp_id"])
@@ -109,7 +115,8 @@ class Agent(LLM, ToolBase):
 
     def _load_tool_obj(self, cpn: dict) -> object:
         from agent.component import component_class
-        param = component_class(cpn["component_name"] + "Param")()
+        tool_name = cpn["component_name"]
+        param = component_class(tool_name + "Param")()
         param.update(cpn["params"])
         try:
             param.check()
@@ -277,19 +284,15 @@ class Agent(LLM, ToolBase):
         else:
             user_request = history[-1]["content"]
 
-        def build_task_desc(prompt: str, user_request: str, tool_metas: list[dict], user_defined_prompt: dict | None = None) -> str:
+        def build_task_desc(prompt: str, user_request: str, user_defined_prompt: dict | None = None) -> str:
             """Build a minimal task_desc by concatenating prompt, query, and tool schemas."""
             user_defined_prompt = user_defined_prompt or {}
 
-            tools_json = json.dumps(tool_metas, ensure_ascii=False, indent=2)
-
             task_desc = (
                 "### Agent Prompt\n"
                 f"{prompt}\n\n"
                 "### User Request\n"
                 f"{user_request}\n\n"
-                "### Tools (schemas)\n"
-                f"{tools_json}\n"
             )
 
             if user_defined_prompt:
@@ -368,7 +371,7 @@ class Agent(LLM, ToolBase):
                 hist.append({"role": "user", "content": content})
 
         st = timer()
-        task_desc = build_task_desc(prompt, user_request, tool_metas, user_defined_prompt)
+        task_desc = build_task_desc(prompt, user_request, user_defined_prompt)
         self.callback("analyze_task", {}, task_desc, elapsed_time=timer()-st)
         for _ in range(self._param.max_rounds + 1):
             if self.check_if_canceled("Agent streaming"):

agent/component/llm.py

@@ -56,7 +56,6 @@ class LLMParam(ComponentParamBase):
         self.check_nonnegative_number(int(self.max_tokens), "[Agent] Max tokens")
         self.check_decimal_float(float(self.top_p), "[Agent] Top P")
         self.check_empty(self.llm_id, "[Agent] LLM")
-        self.check_empty(self.sys_prompt, "[Agent] System prompt")
         self.check_empty(self.prompts, "[Agent] User prompt")
 
     def gen_conf(self):

agent/component/loopitem.py

@@ -113,6 +113,10 @@ class LoopItem(ComponentBase, ABC):
                 return len(var) == 0
             elif operator == "not empty":
                 return len(var) > 0
+        elif var is None:
+            if operator == "empty":
+                return True
+            return False
 
         raise Exception(f"Invalid operator: {operator}")
 

api/apps/document_app.py

@@ -447,6 +447,26 @@ async def metadata_update():
     return get_json_result(data={"updated": updated, "matched_docs": len(target_doc_ids)})
 
 
+@manager.route("/update_metadata_setting", methods=["POST"])  # noqa: F821
+@login_required
+@validate_request("doc_id", "metadata")
+async def update_metadata_setting():
+    req = await get_request_json()
+    if not DocumentService.accessible(req["doc_id"], current_user.id):
+        return get_json_result(data=False, message="No authorization.", code=RetCode.AUTHENTICATION_ERROR)
+
+    e, doc = DocumentService.get_by_id(req["doc_id"])
+    if not e:
+        return get_data_error_result(message="Document not found!")
+
+    DocumentService.update_parser_config(doc.id, {"metadata": req["metadata"]})
+    e, doc = DocumentService.get_by_id(doc.id)
+    if not e:
+        return get_data_error_result(message="Document not found!")
+
+    return get_json_result(data=doc.to_dict())
+
+
 @manager.route("/thumbnails", methods=["GET"])  # noqa: F821
 # @login_required
 def thumbnails():

api/db/services/document_service.py

@@ -33,6 +33,7 @@ from api.db.db_models import DB, Document, Knowledgebase, Task, Tenant, UserTena
 from api.db.db_utils import bulk_insert_into_db
 from api.db.services.common_service import CommonService
 from api.db.services.knowledgebase_service import KnowledgebaseService
+from common.metadata_utils import dedupe_list
 from common.misc_utils import get_uuid
 from common.time_utils import current_timestamp, get_format_time
 from common.constants import LLMType, ParserType, StatusEnum, TaskStatus, SVR_CONSUMER_GROUP_NAME
@@ -696,10 +697,14 @@ class DocumentService(CommonService):
             for k,v in r.meta_fields.items():
                 if k not in meta:
                     meta[k] = {}
-                v = str(v)
-                if v not in meta[k]:
-                    meta[k][v] = []
-                meta[k][v].append(doc_id)
+                if not isinstance(v, list):
+                    v = [v]
+                for vv in v:
+                    if vv not in meta[k]:
+                        if isinstance(vv, list) or isinstance(vv, dict):
+                            continue
+                        meta[k][vv] = []
+                    meta[k][vv].append(doc_id)
         return meta
 
     @classmethod
@@ -797,7 +802,10 @@ class DocumentService(CommonService):
                 match_provided = "match" in upd
                 if isinstance(meta[key], list):
                     if not match_provided:
-                        meta[key] = new_value
+                        if isinstance(new_value, list):
+                            meta[key] = dedupe_list(new_value)
+                        else:
+                            meta[key] = new_value
                         changed = True
                     else:
                         match_value = upd.get("match")
@@ -810,7 +818,7 @@ class DocumentService(CommonService):
                             else:
                                 new_list.append(item)
                         if replaced:
-                            meta[key] = new_list
+                            meta[key] = dedupe_list(new_list)
                             changed = True
                 else:
                     if not match_provided:

api/db/services/knowledgebase_service.py

@@ -411,8 +411,6 @@ class KnowledgebaseService(CommonService):
         ok, _t = TenantService.get_by_id(tenant_id)
         if not ok:
             return False, get_data_error_result(message="Tenant not found.")
-        if kwargs.get("parser_config") and isinstance(kwargs["parser_config"], dict) and not kwargs["parser_config"].get("llm_id"):
-            kwargs["parser_config"]["llm_id"] = _t.llm_id
 
         # Build payload
         kb_id = get_uuid()
@@ -427,6 +425,7 @@ class KnowledgebaseService(CommonService):
 
         # Update parser_config (always override with validated default/merged config)
         payload["parser_config"] = get_parser_config(parser_id, kwargs.get("parser_config"))
+        payload["parser_config"]["llm_id"] = _t.llm_id
 
         return True, payload
 

api/db/services/memory_service.py

@@ -117,6 +117,8 @@ class MemoryService(CommonService):
         if len(memory_name) > MEMORY_NAME_LIMIT:
             return False, f"Memory name {memory_name} exceeds limit of {MEMORY_NAME_LIMIT}."
 
+        timestamp = current_timestamp()
+        format_time = get_format_time()
         # build create dict
         memory_info = {
             "id": get_uuid(),
@@ -126,10 +128,10 @@ class MemoryService(CommonService):
             "embd_id": embd_id,
             "llm_id": llm_id,
             "system_prompt": PromptAssembler.assemble_system_prompt({"memory_type": memory_type}),
-            "create_time": current_timestamp(),
-            "create_date": get_format_time(),
-            "update_time": current_timestamp(),
-            "update_date": get_format_time(),
+            "create_time": timestamp,
+            "create_date": format_time,
+            "update_time": timestamp,
+            "update_date": format_time,
         }
         obj = cls.model(**memory_info).save(force_insert=True)
 

common/metadata_utils.py

@@ -44,21 +44,27 @@ def meta_filter(metas: dict, filters: list[dict], logic: str = "and"):
     def filter_out(v2docs, operator, value):
         ids = []
         for input, docids in v2docs.items():
+
             if operator in ["=", "≠", ">", "<", "≥", "≤"]:
                 try:
+                    if isinstance(input, list):
+                        input = input[0]
                     input = float(input)
                     value = float(value)
                 except Exception:
-                    input = str(input)
-                    value = str(value)
+                    pass
+            if isinstance(input, str):
+                input = input.lower()
+            if isinstance(value, str):
+                value = value.lower()
 
             for conds in [
-                (operator == "contains", str(value).lower() in str(input).lower()),
-                (operator == "not contains", str(value).lower() not in str(input).lower()),
-                (operator == "in", str(input).lower() in str(value).lower()),
-                (operator == "not in", str(input).lower() not in str(value).lower()),
-                (operator == "start with", str(input).lower().startswith(str(value).lower())),
-                (operator == "end with", str(input).lower().endswith(str(value).lower())),
+                (operator == "contains", input in value if not isinstance(input, list) else all([i in value for i in input])),
+                (operator == "not contains", input not in value if not isinstance(input, list) else all([i not in value for i in input])),
+                (operator == "in", input in value if not isinstance(input, list) else all([i in value for i in input])),
+                (operator == "not in", input not in value if not isinstance(input, list) else all([i not in value for i in input])),
+                (operator == "start with", str(input).lower().startswith(str(value).lower())  if not isinstance(input, list) else "".join([str(i).lower() for i in input]).startswith(str(value).lower())),
+                (operator == "end with", str(input).lower().endswith(str(value).lower())  if not isinstance(input, list) else "".join([str(i).lower() for i in input]).endswith(str(value).lower())),
                 (operator == "empty", not input),
                 (operator == "not empty", input),
                 (operator == "=", input == value),
@@ -145,6 +151,18 @@ async def apply_meta_data_filter(
     return doc_ids
 
 
+def dedupe_list(values: list) -> list:
+    seen = set()
+    deduped = []
+    for item in values:
+        key = str(item)
+        if key in seen:
+            continue
+        seen.add(key)
+        deduped.append(item)
+    return deduped
+
+
 def update_metadata_to(metadata, meta):
     if not meta:
         return metadata
@@ -156,11 +174,13 @@ def update_metadata_to(metadata, meta):
             return metadata
     if not isinstance(meta, dict):
         return metadata
+
     for k, v in meta.items():
         if isinstance(v, list):
             v = [vv for vv in v if isinstance(vv, str)]
             if not v:
                 continue
+            v = dedupe_list(v)
         if not isinstance(v, list) and not isinstance(v, str):
             continue
         if k not in metadata:
@@ -171,6 +191,7 @@ def update_metadata_to(metadata, meta):
                 metadata[k].extend(v)
             else:
                 metadata[k].append(v)
+            metadata[k] = dedupe_list(metadata[k])
         else:
             metadata[k] = v
 
@@ -202,4 +223,4 @@ def metadata_schema(metadata: list|None) -> Dict[str, Any]:
     }
 
     json_schema["additionalProperties"] = False
-    return json_schema
+    return json_schema

conf/llm_factories.json

@@ -1258,6 +1258,12 @@
             "status": "1",
             "rank": "810",
             "llm": [
+                {
+                    "llm_name": "MiniMax-M2.1",
+                    "tags": "LLM,CHAT,200k",
+                    "max_tokens": 200000,
+                    "model_type": "chat"
+                },
                 {
                     "llm_name": "MiniMax-M2",
                     "tags": "LLM,CHAT,200k",

deepdoc/parser/pdf_parser.py

@@ -1447,6 +1447,7 @@ class VisionParser(RAGFlowPdfParser):
     def __init__(self, vision_model, *args, **kwargs):
         super().__init__(*args, **kwargs)
         self.vision_model = vision_model
+        self.outlines = []
 
     def __images__(self, fnm, zoomin=3, page_from=0, page_to=299, callback=None):
         try:

docker/docker-compose-base.yml

@@ -72,7 +72,7 @@ services:
   infinity:
     profiles:
       - infinity
-    image: infiniflow/infinity:v0.6.11
+    image: infiniflow/infinity:v0.6.13
     volumes:
       - infinity_data:/var/infinity
       - ./infinity_conf.toml:/infinity_conf.toml

docker/infinity_conf.toml

@@ -1,5 +1,5 @@
 [general]
-version                  = "0.6.11"
+version                  = "0.6.13"
 time_zone                = "utc-8"
 
 [network]

docs/guides/agent/agent_component_reference/http.md

@@ -0,0 +1,90 @@
+---
+sidebar_position: 30
+slug: /http_request_component
+---
+
+# HTTP request component
+
+A component that calls remote services. 
+
+---
+
+An **HTTP request** component lets you access remote APIs or services by providing a URL and an HTTP method, and then receive the response. You can customize headers, parameters, proxies, and timeout settings, and use common methods like GET and POST. It’s useful for exchanging data with external systems in a workflow.
+
+## Prerequisites
+
+- An accessible remote API or service.
+- Add a Token or credentials to the request header, if the target service requires authentication.
+
+## Configurations
+
+### Url
+
+*Required*. The complete request address, for example: http://api.example.com/data.
+
+### Method
+
+The HTTP request method to select. Available options:
+
+- GET
+- POST
+- PUT
+
+### Timeout
+
+The maximum waiting time for the request, in seconds. Defaults to `60`.
+
+### Headers
+
+Custom HTTP headers can be set here, for example:
+
+```http
+{
+  "Accept": "application/json",
+  "Cache-Control": "no-cache",
+  "Connection": "keep-alive"
+}
+```
+
+### Proxy
+
+Optional. The proxy server address to use for this request.
+
+### Clean HTML
+
+`Boolean`: Whether to remove HTML tags from the returned results and keep plain text only.
+
+### Parameter
+
+*Optional*. Parameters to send with the HTTP request. Supports key-value pairs:
+
+- To assign a value using a dynamic system variable, set it as Variable.
+- To override these dynamic values under certain conditions and use a fixed static value instead, Value is the appropriate choice.
+
+
+:::tip NOTE
+- For GET requests, these parameters are appended to the end of the URL.
+- For POST/PUT requests, they are sent as the request body.
+:::
+
+#### Example setting
+
+![](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/http_settings.png)
+
+#### Example response
+
+```html
+{ "args": { "App": "RAGFlow", "Query": "How to do?", "Userid": "241ed25a8e1011f0b979424ebc5b108b" }, "headers": { "Accept": "/", "Accept-Encoding": "gzip, deflate, br, zstd", "Cache-Control": "no-cache", "Host": "httpbin.org", "User-Agent": "python-requests/2.32.2", "X-Amzn-Trace-Id": "Root=1-68c9210c-5aab9088580c130a2f065523" }, "origin": "185.36.193.38", "url": "https://httpbin.org/get?Userid=241ed25a8e1011f0b979424ebc5b108b&App=RAGFlow&Query=How+to+do%3F" }
+```
+
+### Output
+
+The global variable name for the output of the HTTP request component, which can be referenced by other components in the workflow.
+
+- `Result`: `string` The response returned by the remote service.
+
+## Example
+
+This is a usage example: a workflow sends a GET request from the **Begin** component to `https://httpbin.org/get` via the **HTTP Request_0** component, passes parameters to the server, and finally outputs the result through the **Message_0** component.
+
+![](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/http_usage.PNG)

helm/values.yaml

@@ -96,7 +96,7 @@ ragflow:
 infinity:
   image:
     repository: infiniflow/infinity
-    tag: v0.6.11
+    tag: v0.6.13
     pullPolicy: IfNotPresent
     pullSecrets: []
   storage:

pyproject.toml

@@ -46,7 +46,7 @@ dependencies = [
     "groq==0.9.0",
     "grpcio-status==1.67.1",
     "html-text==0.6.2",
-    "infinity-sdk==0.6.11",
+    "infinity-sdk==0.6.13",
     "infinity-emb>=0.0.66,<0.0.67",
     "jira==3.10.5",
     "json-repair==0.35.0",

rag/flow/parser/parser.py

@@ -37,7 +37,6 @@ from rag.app.naive import Docx
 from rag.flow.base import ProcessBase, ProcessParamBase
 from rag.flow.parser.schema import ParserFromUpstream
 from rag.llm.cv_model import Base as VLM
-from rag.nlp import attach_media_context
 from rag.utils.base64_image import image2id
 
 
@@ -86,8 +85,6 @@ class ParserParam(ProcessParamBase):
                     "pdf",
                 ],
                 "output_format": "json",
-                "table_context_size": 0,
-                "image_context_size": 0,
             },
             "spreadsheet": {
                 "parse_method": "deepdoc",  # deepdoc/tcadp_parser
@@ -97,8 +94,6 @@ class ParserParam(ProcessParamBase):
                     "xlsx",
                     "csv",
                 ],
-                "table_context_size": 0,
-                "image_context_size": 0,
             },
             "word": {
                 "suffix": [
@@ -106,14 +101,10 @@ class ParserParam(ProcessParamBase):
                     "docx",
                 ],
                 "output_format": "json",
-                "table_context_size": 0,
-                "image_context_size": 0,
             },
             "text&markdown": {
                 "suffix": ["md", "markdown", "mdx", "txt"],
                 "output_format": "json",
-                "table_context_size": 0,
-                "image_context_size": 0,
             },
             "slides": {
                 "parse_method": "deepdoc",  # deepdoc/tcadp_parser
@@ -122,8 +113,6 @@ class ParserParam(ProcessParamBase):
                     "ppt",
                 ],
                 "output_format": "json",
-                "table_context_size": 0,
-                "image_context_size": 0,
             },
             "image": {
                 "parse_method": "ocr",
@@ -357,11 +346,6 @@ class Parser(ProcessBase):
             elif layout == "table":
                 b["doc_type_kwd"] = "table"
 
-        table_ctx = conf.get("table_context_size", 0) or 0
-        image_ctx = conf.get("image_context_size", 0) or 0
-        if table_ctx or image_ctx:
-            bboxes = attach_media_context(bboxes, table_ctx, image_ctx)
-
         if conf.get("output_format") == "json":
             self.set_output("json", bboxes)
         if conf.get("output_format") == "markdown":
@@ -436,11 +420,6 @@ class Parser(ProcessBase):
                     if table:
                         result.append({"text": table, "doc_type_kwd": "table"})
 
-                table_ctx = conf.get("table_context_size", 0) or 0
-                image_ctx = conf.get("image_context_size", 0) or 0
-                if table_ctx or image_ctx:
-                    result = attach_media_context(result, table_ctx, image_ctx)
-
                 self.set_output("json", result)
 
             elif output_format == "markdown":
@@ -476,11 +455,6 @@ class Parser(ProcessBase):
             sections = [{"text": section[0], "image": section[1]} for section in sections if section]
             sections.extend([{"text": tb, "image": None, "doc_type_kwd": "table"} for ((_, tb), _) in tbls])
 
-            table_ctx = conf.get("table_context_size", 0) or 0
-            image_ctx = conf.get("image_context_size", 0) or 0
-            if table_ctx or image_ctx:
-                sections = attach_media_context(sections, table_ctx, image_ctx)
-
             self.set_output("json", sections)
         elif conf.get("output_format") == "markdown":
             markdown_text = docx_parser.to_markdown(name, binary=blob)
@@ -536,11 +510,6 @@ class Parser(ProcessBase):
                     if table:
                         result.append({"text": table, "doc_type_kwd": "table"})
 
-                table_ctx = conf.get("table_context_size", 0) or 0
-                image_ctx = conf.get("image_context_size", 0) or 0
-                if table_ctx or image_ctx:
-                    result = attach_media_context(result, table_ctx, image_ctx)
-
                 self.set_output("json", result)
         else:
             # Default DeepDOC parser (supports .pptx format)
@@ -554,10 +523,6 @@ class Parser(ProcessBase):
             # json
             assert conf.get("output_format") == "json", "have to be json for ppt"
             if conf.get("output_format") == "json":
-                table_ctx = conf.get("table_context_size", 0) or 0
-                image_ctx = conf.get("image_context_size", 0) or 0
-                if table_ctx or image_ctx:
-                    sections = attach_media_context(sections, table_ctx, image_ctx)
                 self.set_output("json", sections)
 
     def _markdown(self, name, blob):
@@ -597,11 +562,6 @@ class Parser(ProcessBase):
 
                 json_results.append(json_result)
 
-            table_ctx = conf.get("table_context_size", 0) or 0
-            image_ctx = conf.get("image_context_size", 0) or 0
-            if table_ctx or image_ctx:
-                json_results = attach_media_context(json_results, table_ctx, image_ctx)
-
             self.set_output("json", json_results)
         else:
             self.set_output("text", "\n".join([section_text for section_text, _ in sections]))

rag/flow/splitter/splitter.py

@@ -23,7 +23,7 @@ from rag.utils.base64_image import id2image, image2id
 from deepdoc.parser.pdf_parser import RAGFlowPdfParser
 from rag.flow.base import ProcessBase, ProcessParamBase
 from rag.flow.splitter.schema import SplitterFromUpstream
-from rag.nlp import naive_merge, naive_merge_with_images
+from rag.nlp import attach_media_context, naive_merge, naive_merge_with_images
 from common import settings
 
 
@@ -34,11 +34,15 @@ class SplitterParam(ProcessParamBase):
         self.delimiters = ["\n"]
         self.overlapped_percent = 0
         self.children_delimiters = []
+        self.table_context_size = 0
+        self.image_context_size = 0
 
     def check(self):
         self.check_empty(self.delimiters, "Delimiters.")
         self.check_positive_integer(self.chunk_token_size, "Chunk token size.")
         self.check_decimal_float(self.overlapped_percent, "Overlapped percentage: [0, 1)")
+        self.check_nonnegative_number(self.table_context_size, "Table context size.")
+        self.check_nonnegative_number(self.image_context_size, "Image context size.")
 
     def get_input_form(self) -> dict[str, dict]:
         return {}
@@ -103,8 +107,18 @@ class Splitter(ProcessBase):
             return
 
         # json
+        json_result = from_upstream.json_result or []
+        if self._param.table_context_size or self._param.image_context_size:
+            for ck in json_result:
+                if "image" not in ck and ck.get("img_id") and not (isinstance(ck.get("text"), str) and ck.get("text").strip()):
+                    ck["image"] = True
+            attach_media_context(json_result, self._param.table_context_size, self._param.image_context_size)
+            for ck in json_result:
+                if ck.get("image") is True:
+                    del ck["image"]
+
         sections, section_images = [], []
-        for o in from_upstream.json_result or []:
+        for o in json_result:
             sections.append((o.get("text", ""), o.get("position_tag", "")))
             section_images.append(id2image(o.get("img_id"), partial(settings.STORAGE_IMPL.get, tenant_id=self._canvas._tenant_id)))
 

rag/flow/tokenizer/tokenizer.py

@@ -105,6 +105,9 @@ class Tokenizer(ProcessBase):
 
     async def _invoke(self, **kwargs):
         try:
+            chunks = kwargs.get("chunks")
+            kwargs["chunks"] = [c for c in chunks if c is not None]
+
             from_upstream = TokenizerFromUpstream.model_validate(kwargs)
         except Exception as e:
             self.set_output("_ERROR", f"Input error: {str(e)}")

rag/nlp/__init__.py

@@ -348,7 +348,8 @@ def tokenize_table(tbls, doc, eng, batch_size=10):
             d["doc_type_kwd"] = "table"
             if img:
                 d["image"] = img
-                d["doc_type_kwd"] = "image"
+                if d["content_with_weight"].find("<tr>") < 0:
+                    d["doc_type_kwd"] = "image"
             if poss:
                 add_positions(d, poss)
             res.append(d)
@@ -361,7 +362,8 @@ def tokenize_table(tbls, doc, eng, batch_size=10):
             d["doc_type_kwd"] = "table"
             if img:
                 d["image"] = img
-                d["doc_type_kwd"] = "image"
+                if d["content_with_weight"].find("<tr>") < 0:
+                    d["doc_type_kwd"] = "image"
             add_positions(d, poss)
             res.append(d)
     return res

rag/prompts/generator.py

@@ -339,7 +339,7 @@ def tool_schema(tools_description: list[dict], complete_task=False):
         }
     for idx, tool in enumerate(tools_description):
         name = tool["function"]["name"]
-        desc[f"{name}_{idx}"] = tool
+        desc[name] = tool
 
     return "\n\n".join([f"## {i+1}. {fnm}\n{json.dumps(des, ensure_ascii=False, indent=4)}" for i, (fnm, des) in enumerate(desc.items())])
 

rag/prompts/next_step.md

@@ -94,7 +94,7 @@ This content will NOT be shown to the user.
 ## Step 2: Structured Reflection (MANDATORY before `complete_task`)
 
 ### Context
-- Goal: {{ task_analysis }}
+- Goal: Reflect on the current task based on the full conversation context
 - Executed tool calls so far (if any): reflect from conversation history
 
 ### Task Complexity Assessment

rag/svr/task_executor.py

@@ -395,9 +395,9 @@ async def build_chunks(task, progress_callback):
             await asyncio.gather(*tasks, return_exceptions=True)
             raise
         metadata = {}
-        for ck in cks:
-            metadata = update_metadata_to(metadata, ck["metadata_obj"])
-            del ck["metadata_obj"]
+        for doc in docs:
+            metadata = update_metadata_to(metadata, doc["metadata_obj"])
+            del doc["metadata_obj"]
         if metadata:
             e, doc = DocumentService.get_by_id(task["doc_id"])
             if e:

rag/utils/base64_image.py

@@ -82,7 +82,7 @@ def id2image(image_id:str|None, storage_get_func: partial):
         return
     bkt, nm = image_id.split("-")
     try:
-        blob = storage_get_func(bucket=bkt, filename=nm)
+        blob = storage_get_func(bucket=bkt, fnm=nm)
         if not blob:
             return
         return Image.open(BytesIO(blob))

uv.lock

@@ -3051,7 +3051,7 @@ wheels = [
 
 [[package]]
 name = "infinity-sdk"
-version = "0.6.11"
+version = "0.6.13"
 source = { registry = "https://pypi.tuna.tsinghua.edu.cn/simple" }
 dependencies = [
     { name = "datrie" },
@@ -3068,9 +3068,9 @@ dependencies = [
     { name = "sqlglot", extra = ["rs"] },
     { name = "thrift" },
 ]
-sdist = { url = "https://pypi.tuna.tsinghua.edu.cn/packages/6e/6d/294b4c8fb36f874c92576107fab22da6d64f567fd3e24a312d7bcba5f17a/infinity_sdk-0.6.11.tar.gz", hash = "sha256:f78acd5439c3837715ab308c49be04b416bdaa42a5f4fb840682639ee39d435f", size = 29518792, upload-time = "2025-12-08T05:58:35.167Z" }
+sdist = { url = "https://pypi.tuna.tsinghua.edu.cn/packages/03/de/56fdc0fa962d5a8e0aa68d16f5321b2d88d79fceb7d0d6cfdde338b65d05/infinity_sdk-0.6.13.tar.gz", hash = "sha256:faf7bc23de7fa549a3842753eddad54ae551ada9df4fff25421658a7fa6fa8c2", size = 29518902, upload-time = "2025-12-24T10:00:01.483Z" }
 wheels = [
-    { url = "https://pypi.tuna.tsinghua.edu.cn/packages/18/cf/d7d1bc584c8f7d1cd8f75c39067179b98ca4bcbe5a86c61e7dbc2b8e692d/infinity_sdk-0.6.11-py3-none-any.whl", hash = "sha256:ec5ac3e710f29db4b875d3e24e20e391ad64270a5e5d189295cc91c362af74d1", size = 29737403, upload-time = "2025-12-08T05:54:58.798Z" },
+    { url = "https://pypi.tuna.tsinghua.edu.cn/packages/f4/a0/8f1e134fdf4ca8bebac7b62caace1816953bb5ffc720d9f0004246c8c38d/infinity_sdk-0.6.13-py3-none-any.whl", hash = "sha256:c08a523d2c27e9a7e6e88be640970530b4661a67c3e9dc3e1aa89533a822fd78", size = 29737403, upload-time = "2025-12-24T09:56:16.93Z" },
 ]
 
 [[package]]
@@ -6229,7 +6229,7 @@ requires-dist = [
     { name = "grpcio-status", specifier = "==1.67.1" },
     { name = "html-text", specifier = "==0.6.2" },
     { name = "infinity-emb", specifier = ">=0.0.66,<0.0.67" },
-    { name = "infinity-sdk", specifier = "==0.6.11" },
+    { name = "infinity-sdk", specifier = "==0.6.13" },
     { name = "jira", specifier = "==3.10.5" },
     { name = "json-repair", specifier = "==0.35.0" },
     { name = "langfuse", specifier = ">=2.60.0" },

web/src/components/chunk-method-dialog/index.tsx

@@ -18,7 +18,9 @@ import { useFetchKnowledgeBaseConfiguration } from '@/hooks/use-knowledge-reques
 import { IModalProps } from '@/interfaces/common';
 import { IParserConfig } from '@/interfaces/database/document';
 import { IChangeParserConfigRequestBody } from '@/interfaces/request/document';
+import { MetadataType } from '@/pages/dataset/components/metedata/hooks/use-manage-modal';
 import {
+  AutoMetadata,
   ChunkMethodItem,
   EnableTocToggle,
   ImageContextWindow,
@@ -86,6 +88,7 @@ export function ChunkMethodDialog({
   visible,
   parserConfig,
   loading,
+  documentId,
 }: IProps) {
   const { t } = useTranslation();
 
@@ -142,6 +145,18 @@ export function ChunkMethodDialog({
         pages: z
           .array(z.object({ from: z.coerce.number(), to: z.coerce.number() }))
           .optional(),
+        metadata: z
+          .array(
+            z
+              .object({
+                key: z.string().optional(),
+                description: z.string().optional(),
+                enum: z.array(z.string().optional()).optional(),
+              })
+              .optional(),
+          )
+          .optional(),
+        enable_metadata: z.boolean().optional(),
       }),
     })
     .superRefine((data, ctx) => {
@@ -373,6 +388,10 @@ export function ChunkMethodDialog({
                   )}
                   {showAutoKeywords(selectedTag) && (
                     <>
+                      <AutoMetadata
+                        type={MetadataType.SingleFileSetting}
+                        otherData={{ documentId }}
+                      />
                       <AutoKeywordsFormField></AutoKeywordsFormField>
                       <AutoQuestionsFormField></AutoQuestionsFormField>
                     </>

web/src/components/chunk-method-dialog/use-default-parser-values.ts

@@ -36,9 +36,11 @@ export function useDefaultParserValues() {
       // },
       entity_types: [],
       pages: [],
+      metadata: [],
+      enable_metadata: false,
     };
 
-    return defaultParserValues;
+    return defaultParserValues as IParserConfig;
   }, [t]);
 
   return defaultParserValues;

web/src/components/dynamic-form.tsx

@@ -35,6 +35,7 @@ import { cn } from '@/lib/utils';
 import { t } from 'i18next';
 import { Loader } from 'lucide-react';
 import { MultiSelect, MultiSelectOptionType } from './ui/multi-select';
+import { Switch } from './ui/switch';
 
 // Field type enumeration
 export enum FormFieldType {
@@ -46,6 +47,7 @@ export enum FormFieldType {
   Select = 'select',
   MultiSelect = 'multi-select',
   Checkbox = 'checkbox',
+  Switch = 'switch',
   Tag = 'tag',
   Custom = 'custom',
 }
@@ -154,6 +156,7 @@ export const generateSchema = (fields: FormFieldConfig[]): ZodSchema<any> => {
           }
           break;
         case FormFieldType.Checkbox:
+        case FormFieldType.Switch:
           fieldSchema = z.boolean();
           break;
         case FormFieldType.Tag:
@@ -193,6 +196,8 @@ export const generateSchema = (fields: FormFieldConfig[]): ZodSchema<any> => {
     if (
       field.type !== FormFieldType.Number &&
       field.type !== FormFieldType.Checkbox &&
+      field.type !== FormFieldType.Switch &&
+      field.type !== FormFieldType.Custom &&
       field.type !== FormFieldType.Tag &&
       field.required
     ) {
@@ -289,7 +294,10 @@ const generateDefaultValues = <T extends FieldValues>(
       const lastKey = keys[keys.length - 1];
       if (field.defaultValue !== undefined) {
         current[lastKey] = field.defaultValue;
-      } else if (field.type === FormFieldType.Checkbox) {
+      } else if (
+        field.type === FormFieldType.Checkbox ||
+        field.type === FormFieldType.Switch
+      ) {
         current[lastKey] = false;
       } else if (field.type === FormFieldType.Tag) {
         current[lastKey] = [];
@@ -299,7 +307,10 @@ const generateDefaultValues = <T extends FieldValues>(
     } else {
       if (field.defaultValue !== undefined) {
         defaultValues[field.name] = field.defaultValue;
-      } else if (field.type === FormFieldType.Checkbox) {
+      } else if (
+        field.type === FormFieldType.Checkbox ||
+        field.type === FormFieldType.Switch
+      ) {
         defaultValues[field.name] = false;
       } else if (
         field.type === FormFieldType.Tag ||
@@ -502,6 +513,32 @@ export const RenderField = ({
           )}
         />
       );
+    case FormFieldType.Switch:
+      return (
+        <RAGFlowFormItem
+          {...field}
+          labelClassName={labelClassName || field.labelClassName}
+        >
+          {(fieldProps) => {
+            const finalFieldProps = field.onChange
+              ? {
+                  ...fieldProps,
+                  onChange: (checked: boolean) => {
+                    fieldProps.onChange(checked);
+                    field.onChange?.(checked);
+                  },
+                }
+              : fieldProps;
+            return (
+              <Switch
+                checked={finalFieldProps.value as boolean}
+                onCheckedChange={(checked) => finalFieldProps.onChange(checked)}
+                disabled={field.disabled}
+              />
+            );
+          }}
+        </RAGFlowFormItem>
+      );
 
     case FormFieldType.Tag:
       return (

web/src/components/file-uploader.tsx

@@ -15,6 +15,7 @@ import { Progress } from '@/components/ui/progress';
 import { useControllableState } from '@/hooks/use-controllable-state';
 import { cn, formatBytes } from '@/lib/utils';
 import { useTranslation } from 'react-i18next';
+import { Tooltip, TooltipContent, TooltipTrigger } from './ui/tooltip';
 
 function isFileWithPreview(file: File): file is File & { preview: string } {
   return 'preview' in file && typeof file.preview === 'string';
@@ -58,10 +59,17 @@ function FileCard({ file, progress, onRemove }: FileCardProps) {
         </div>
         <div className="flex flex-col flex-1 gap-2 overflow-hidden">
           <div className="flex flex-col gap-px">
-            <p className="line-clamp-1 text-sm font-medium text-foreground/80 text-ellipsis">
-              {file.name}
-            </p>
-            <p className="text-xs text-muted-foreground">
+            <Tooltip>
+              <TooltipTrigger asChild>
+                <p className=" w-fit line-clamp-1 text-sm font-medium text-foreground/80 text-ellipsis truncate max-w-[370px]">
+                  {file.name}
+                </p>
+              </TooltipTrigger>
+              <TooltipContent className="border border-border-button">
+                {file.name}
+              </TooltipContent>
+            </Tooltip>
+            <p className="text-xs text-text-secondary">
               {formatBytes(file.size)}
             </p>
           </div>
@@ -311,7 +319,7 @@ export function FileUploader(props: FileUploaderProps) {
                     />
                   </div>
                   <div className="flex flex-col gap-px">
-                    <p className="font-medium text-text-secondary">
+                    <p className="font-medium text-text-secondary ">
                       {title || t('knowledgeDetails.uploadTitle')}
                     </p>
                     <p className="text-sm text-text-disabled">

web/src/components/list-filter-bar/filter-field.tsx

@@ -31,14 +31,16 @@ const handleCheckChange = ({
           (value: string) => value !== item.id.toString(),
         );
 
-    const newValue = {
-      ...currentValue,
-      [parentId]: newParentValues,
-    };
+    const newValue = newParentValues?.length
+      ? {
+          ...currentValue,
+          [parentId]: newParentValues,
+        }
+      : { ...currentValue };
 
-    if (newValue[parentId].length === 0) {
-      delete newValue[parentId];
-    }
+    // if (newValue[parentId].length === 0) {
+    //   delete newValue[parentId];
+    // }
 
     return field.onChange(newValue);
   } else {
@@ -66,20 +68,31 @@ const FilterItem = memo(
   }) => {
     return (
       <div
-        className={`flex items-center justify-between text-text-primary text-xs ${level > 0 ? 'ml-4' : ''}`}
+        className={`flex items-center justify-between text-text-primary text-xs ${level > 0 ? 'ml-1' : ''}`}
       >
-        <FormItem className="flex flex-row space-x-3 space-y-0 items-center">
+        <FormItem className="flex flex-row space-x-3 space-y-0 items-center ">
           <FormControl>
-            <Checkbox
-              checked={field.value?.includes(item.id.toString())}
-              onCheckedChange={(checked: boolean) =>
-                handleCheckChange({ checked, field, item })
-              }
-            />
+            <div className="flex space-x-3">
+              <Checkbox
+                checked={field.value?.includes(item.id.toString())}
+                onCheckedChange={(checked: boolean) =>
+                  handleCheckChange({ checked, field, item })
+                }
+                // className="hidden group-hover:block"
+              />
+              <FormLabel
+                onClick={() =>
+                  handleCheckChange({
+                    checked: !field.value?.includes(item.id.toString()),
+                    field,
+                    item,
+                  })
+                }
+              >
+                {item.label}
+              </FormLabel>
+            </div>
           </FormControl>
-          <FormLabel onClick={(e) => e.stopPropagation()}>
-            {item.label}
-          </FormLabel>
         </FormItem>
         {item.count !== undefined && (
           <span className="text-sm">{item.count}</span>
@@ -107,11 +120,11 @@ export const FilterField = memo(
       <FormField
         key={item.id}
         control={form.control}
-        name={parent.field as string}
+        name={parent.field?.toString() as string}
         render={({ field }) => {
           if (hasNestedList) {
             return (
-              <div className={`flex flex-col gap-2 ${level > 0 ? 'ml-4' : ''}`}>
+              <div className={`flex flex-col gap-2 ${level > 0 ? 'ml-1' : ''}`}>
                 <div
                   className="flex items-center justify-between cursor-pointer"
                   onClick={() => {
@@ -138,23 +151,6 @@ export const FilterField = memo(
                       }}
                       level={level + 1}
                     />
-                    // <FilterItem key={child.id} item={child} field={child.field} level={level+1} />
-                    // <div
-                    //   className="flex flex-row space-x-3 space-y-0 items-center"
-                    //   key={child.id}
-                    // >
-                    //   <FormControl>
-                    //     <Checkbox
-                    //       checked={field.value?.includes(child.id.toString())}
-                    //       onCheckedChange={(checked) =>
-                    //         handleCheckChange({ checked, field, item: child })
-                    //       }
-                    //     />
-                    //   </FormControl>
-                    //   <FormLabel onClick={(e) => e.stopPropagation()}>
-                    //     {child.label}
-                    //   </FormLabel>
-                    // </div>
                   ))}
               </div>
             );

web/src/components/list-filter-bar/filter-popover.tsx

@@ -11,8 +11,8 @@ import {
   useMemo,
   useState,
 } from 'react';
-import { useForm } from 'react-hook-form';
-import { ZodArray, ZodString, z } from 'zod';
+import { FieldPath, useForm } from 'react-hook-form';
+import { z } from 'zod';
 
 import { Button } from '@/components/ui/button';
 
@@ -71,34 +71,37 @@ function CheckboxFormMultiple({
     }, {});
   }, [resolvedFilters]);
 
+  // const FormSchema = useMemo(() => {
+  //   if (resolvedFilters.length === 0) {
+  //     return z.object({});
+  //   }
+
+  //   return z.object(
+  //     resolvedFilters.reduce<
+  //       Record<
+  //         string,
+  //         ZodArray<ZodString, 'many'> | z.ZodObject<any> | z.ZodOptional<any>
+  //       >
+  //     >((pre, cur) => {
+  //       const hasNested = cur.list?.some(
+  //         (item) => item.list && item.list.length > 0,
+  //       );
+
+  //       if (hasNested) {
+  //         pre[cur.field] = z
+  //           .record(z.string(), z.array(z.string().optional()).optional())
+  //           .optional();
+  //       } else {
+  //         pre[cur.field] = z.array(z.string().optional()).optional();
+  //       }
+
+  //       return pre;
+  //     }, {}),
+  //   );
+  // }, [resolvedFilters]);
   const FormSchema = useMemo(() => {
-    if (resolvedFilters.length === 0) {
-      return z.object({});
-    }
-
-    return z.object(
-      resolvedFilters.reduce<
-        Record<
-          string,
-          ZodArray<ZodString, 'many'> | z.ZodObject<any> | z.ZodOptional<any>
-        >
-      >((pre, cur) => {
-        const hasNested = cur.list?.some(
-          (item) => item.list && item.list.length > 0,
-        );
-
-        if (hasNested) {
-          pre[cur.field] = z
-            .record(z.string(), z.array(z.string().optional()).optional())
-            .optional();
-        } else {
-          pre[cur.field] = z.array(z.string().optional()).optional();
-        }
-
-        return pre;
-      }, {}),
-    );
-  }, [resolvedFilters]);
+    return z.object({});
+  }, []);
 
   const form = useForm<z.infer<typeof FormSchema>>({
     resolver: resolvedFilters.length > 0 ? zodResolver(FormSchema) : undefined,
@@ -178,7 +181,9 @@ function CheckboxFormMultiple({
                 <FormField
                   key={x.field}
                   control={form.control}
-                  name={x.field}
+                  name={
+                    x.field.toString() as FieldPath<z.infer<typeof FormSchema>>
+                  }
                   render={() => (
                     <FormItem className="space-y-4">
                       <div>
@@ -186,19 +191,20 @@ function CheckboxFormMultiple({
                           {x.label}
                         </FormLabel>
                       </div>
-                      {x.list.map((item) => {
-                        return (
-                          <FilterField
-                            key={item.id}
-                            item={{ ...item }}
-                            parent={{
-                              ...x,
-                              id: x.field,
-                              // field: `${x.field}${item.field ? '.' + item.field : ''}`,
-                            }}
-                          />
-                        );
-                      })}
+                      {x.list?.length &&
+                        x.list.map((item) => {
+                          return (
+                            <FilterField
+                              key={item.id}
+                              item={{ ...item }}
+                              parent={{
+                                ...x,
+                                id: x.field,
+                                // field: `${x.field}${item.field ? '.' + item.field : ''}`,
+                              }}
+                            />
+                          );
+                        })}
                       <FormMessage />
                     </FormItem>
                   )}

web/src/hooks/use-document-request.ts

@@ -14,6 +14,7 @@ import {
   IDocumentMetaRequestBody,
 } from '@/interfaces/request/document';
 import i18n from '@/locales/config';
+import { EMPTY_METADATA_FIELD } from '@/pages/dataset/dataset/use-select-filters';
 import kbService, { listDocument } from '@/services/knowledge-service';
 import api, { api_host } from '@/utils/api';
 import { buildChunkHighlights } from '@/utils/document-util';
@@ -114,6 +115,20 @@ export const useFetchDocumentList = () => {
     refetchInterval: isLoop ? 5000 : false,
     enabled: !!knowledgeId || !!id,
     queryFn: async () => {
+      let run = [] as any;
+      let returnEmptyMetadata = false;
+      if (filterValue.run && Array.isArray(filterValue.run)) {
+        run = [...(filterValue.run as string[])];
+        const returnEmptyMetadataIndex = run.findIndex(
+          (r: string) => r === EMPTY_METADATA_FIELD,
+        );
+        if (returnEmptyMetadataIndex > -1) {
+          returnEmptyMetadata = true;
+          run.splice(returnEmptyMetadataIndex, 1);
+        }
+      } else {
+        run = filterValue.run;
+      }
       const ret = await listDocument(
         {
           kb_id: knowledgeId || id,
@@ -123,7 +138,8 @@ export const useFetchDocumentList = () => {
         },
         {
           suffix: filterValue.type as string[],
-          run_status: filterValue.run as string[],
+          run_status: run as string[],
+          return_empty_metadata: returnEmptyMetadata,
           metadata: filterValue.metadata as Record<string, string[]>,
         },
       );

web/src/interfaces/database/document.ts

@@ -43,6 +43,18 @@ export interface IParserConfig {
   task_page_size?: number;
   raptor?: Raptor;
   graphrag?: GraphRag;
+  image_context_window?: number;
+  mineru_parse_method?: 'auto' | 'txt' | 'ocr';
+  mineru_formula_enable?: boolean;
+  mineru_table_enable?: boolean;
+  mineru_lang?: string;
+  entity_types?: string[];
+  metadata?: Array<{
+    key?: string;
+    description?: string;
+    enum?: string[];
+  }>;
+  enable_metadata?: boolean;
 }
 
 interface Raptor {

web/src/interfaces/request/knowledge.ts

@@ -33,5 +33,6 @@ export interface IFetchKnowledgeListRequestParams {
 export interface IFetchDocumentListRequestBody {
   suffix?: string[];
   run_status?: string[];
+  return_empty_metadata?: boolean;
   metadata?: Record<string, string[]>;
 }

web/src/locales/en.ts

@@ -176,6 +176,15 @@ Procedural Memory: Learned skills, habits, and automated procedures.`,
     },
     knowledgeDetails: {
       metadata: {
+        descriptionTip:
+          'Provide descriptions or examples to guide LLM extract values for this field. If left empty, it will rely on the field name.',
+        restrictTDefinedValuesTip:
+          'Enum Mode: Restricts LLM extraction to match preset values only. Define values below.',
+        valueExists:
+          'Value already exists. Confirm to merge duplicates and combine all associated files.',
+        fieldNameExists:
+          'Field name already exists. Confirm to merge duplicates and combine all associated files.',
+        fieldExists: 'Field already exists.',
         fieldSetting: 'Field settings',
         changesAffectNewParses: 'Changes affect new parses only.',
         editMetadataForDataset: 'View and edit metadata for ',
@@ -185,12 +194,25 @@ Procedural Memory: Learned skills, habits, and automated procedures.`,
         manageMetadata: 'Manage metadata',
         metadata: 'Metadata',
         values: 'Values',
+        value: 'Value',
         action: 'Action',
         field: 'Field',
         description: 'Description',
         fieldName: 'Field name',
         editMetadata: 'Edit metadata',
+        deleteWarn: 'This {{field}} will be removed from all associated files',
+        deleteManageFieldAllWarn:
+          'This field and all its corresponding values will be deleted from all associated files.',
+        deleteManageValueAllWarn:
+          'This value will be deleted from from all associated files.',
+        deleteManageFieldSingleWarn:
+          'This field and all its corresponding values will be deleted from this files.',
+        deleteManageValueSingleWarn:
+          'This value will be deleted from this files.',
+        deleteSettingFieldWarn: `This field will be deleted; existing metadata won't be affected.`,
+        deleteSettingValueWarn: `This value will be deleted; existing metadata won't be affected.`,
       },
+      emptyMetadata: 'No metadata',
       metadataField: 'Metadata field',
       systemAttribute: 'System attribute',
       localUpload: 'Local upload',
@@ -334,9 +356,9 @@ Procedural Memory: Learned skills, habits, and automated procedures.`,
       html4excel: 'Excel to HTML',
       html4excelTip: `Use with the General chunking method. When disabled, spreadsheets (XLSX or XLS(Excel 97-2003)) in the knowledge base will be parsed into key-value pairs. When enabled, they will be parsed into HTML tables, splitting every 12 rows if the original table has more than 12 rows. See https://ragflow.io/docs/dev/enable_excel2html for details.`,
       autoKeywords: 'Auto-keyword',
-      autoKeywordsTip: `Automatically extract N keywords for each chunk to increase their ranking for queries containing those keywords. Be aware that extra tokens will be consumed by the chat model specified in 'System model settings'. You can check or update the added keywords for a chunk from the chunk list. For details, see https://ragflow.io/docs/dev/autokeyword_autoquestion.`,
+      autoKeywordsTip: `Automatically extract N keywords for each chunk to increase their ranking for queries containing those keywords. Be aware that extra tokens will be consumed by the indexing model specified in 'Configuration'. You can check or update the added keywords for a chunk from the chunk list. For details, see https://ragflow.io/docs/dev/autokeyword_autoquestion.`,
       autoQuestions: 'Auto-question',
-      autoQuestionsTip: `Automatically extract N questions for each chunk to increase their ranking for queries containing those questions. You can check or update the added questions for a chunk from the chunk list. This feature will not disrupt the chunking process if an error occurs, except that it may add an empty result to the original chunk. Be aware that extra tokens will be consumed by the LLM specified in 'System model settings'. For details, see https://ragflow.io/docs/dev/autokeyword_autoquestion.`,
+      autoQuestionsTip: `Automatically extract N questions for each chunk to increase their ranking for queries containing those questions. You can check or update the added questions for a chunk from the chunk list. This feature will not disrupt the chunking process if an error occurs, except that it may add an empty result to the original chunk. Be aware that extra tokens will be consumed by the indexing model specified in 'Configuration'. For details, see https://ragflow.io/docs/dev/autokeyword_autoquestion.`,
       redo: 'Do you want to clear the existing {{chunkNum}} chunks?',
       setMetaData: 'Set meta data',
       pleaseInputJson: 'Please enter JSON',
@@ -1607,6 +1629,8 @@ Example: Virtual Hosted Style`,
         notEmpty: 'Not empty',
         in: 'In',
         notIn: 'Not in',
+        is: 'Is',
+        isNot: 'Is not',
       },
       switchLogicOperatorOptions: {
         and: 'AND',
@@ -1868,6 +1892,8 @@ This process aggregates variables from multiple branches into a single variable
       beginInputTip:
         'By defining input parameters, this content can be accessed by other components in subsequent processes.',
       query: 'Query variables',
+      switchPromptMessage:
+        'The prompt words will change. Please confirm whether you want to discard the existing prompt words?',
       queryRequired: 'Query is required',
       queryTip: 'Select the variable you want to use',
       agent: 'Agent',

web/src/locales/zh.ts

@@ -168,6 +168,10 @@ export default {
     },
     knowledgeDetails: {
       metadata: {
+        descriptionTip:
+          '提供描述或示例来指导大语言模型为此字段提取值。如果留空，将依赖字段名称。',
+        restrictTDefinedValuesTip:
+          '枚举模式：限制大语言模型仅提取预设值。在下方定义值。',
         fieldSetting: '字段设置',
         changesAffectNewParses: '更改仅影响新解析。',
         editMetadataForDataset: '查看和编辑元数据于 ',
@@ -177,12 +181,25 @@ export default {
         manageMetadata: '管理元数据',
         metadata: '元数据',
         values: '值',
+        value: '值',
         action: '操作',
         field: '字段',
         description: '描述',
         fieldName: '字段名',
         editMetadata: '编辑元数据',
+        valueExists: '值已存在。确认合并重复项并组合所有关联文件。',
+        fieldNameExists: '字段名已存在。确认合并重复项并组合所有关联文件。',
+        fieldExists: '字段名已存在。',
+        deleteWarn: '此 {{field}} 将从所有关联文件中移除',
+        deleteManageFieldAllWarn:
+          '此字段及其所有对应值将从所有关联的文件中删除。',
+        deleteManageValueAllWarn: '此值将从所有关联的文件中删除。',
+        deleteManageFieldSingleWarn: '此字段及其所有对应值将从此文件中删除。',
+        deleteManageValueSingleWarn: '此值将从此文件中删除。',
+        deleteSettingFieldWarn: `此字段将被删除；现有元数据不会受到影响。`,
+        deleteSettingValueWarn: `此值将被删除；现有元数据不会受到影响。`,
       },
+      emptyMetadata: '无元数据',
       localUpload: '本地上传',
       fileSize: '文件大小',
       fileType: '文件类型',
@@ -311,9 +328,9 @@ export default {
       html4excel: '表格转HTML',
       html4excelTip: `与 General 切片方法配合使用。未开启状态下，表格文件（XLSX、XLS（Excel 97-2003））会按行解析为键值对。开启后，表格文件会被解析为 HTML 表格。若原始表格超过 12 行，系统会自动按每 12 行拆分为多个 HTML 表格。欲了解更多详情，请参阅 https://ragflow.io/docs/dev/enable_excel2html。`,
       autoKeywords: '自动关键词提取',
-      autoKeywordsTip: `自动为每个文本块中提取 N 个关键词，用以提升查询精度。请注意：该功能采用“系统模型设置”中设置的默认聊天模型提取关键词，因此也会产生更多 Token 消耗。另外，你也可以手动更新生成的关键词。详情请见 https://ragflow.io/docs/dev/autokeyword_autoquestion。`,
+      autoKeywordsTip: `自动为每个文本块中提取 N 个关键词，用以提升查询精度。请注意：该功能采用在“配置”中指定的索引模型提取关键词，因此也会产生更多 Token 消耗。另外，你也可以手动更新生成的关键词。详情请见 https://ragflow.io/docs/dev/autokeyword_autoquestion。`,
       autoQuestions: '自动问题提取',
-      autoQuestionsTip: `利用“系统模型设置”中设置的 chat model 对知识库的每个文本块提取 N 个问题以提高其排名得分。请注意，开启后将消耗额外的 token。您可以在块列表中查看、编辑结果。如果自动问题提取发生错误，不会妨碍整个分块过程，只会将空结果添加到原始文本块。详情请见 https://ragflow.io/docs/dev/autokeyword_autoquestion。`,
+      autoQuestionsTip: `利用在“配置”中指定的索引模型 对知识库的每个文本块提取 N 个问题以提高其排名得分。请注意，开启后将消耗额外的 token。您可以在块列表中查看、编辑结果。如果自动问题提取发生错误，不会妨碍整个分块过程，只会将空结果添加到原始文本块。详情请见 https://ragflow.io/docs/dev/autokeyword_autoquestion。`,
       redo: '是否清空已有 {{chunkNum}}个 chunk？',
       setMetaData: '设置元数据',
       pleaseInputJson: '请输入JSON',
@@ -1505,6 +1522,8 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
         endWith: '结束是',
         empty: '为空',
         notEmpty: '不为空',
+        is: '是',
+        isNot: '不是',
       },
       switchLogicOperatorOptions: {
         and: '与',

web/src/pages/agent/chat/box.tsx

@@ -4,6 +4,7 @@ import { useSendAgentMessage } from './use-send-agent-message';
 
 import { FileUploadProps } from '@/components/file-upload';
 import { NextMessageInput } from '@/components/message-input/next';
+import MarkdownContent from '@/components/next-markdown-content';
 import MessageItem from '@/components/next-message-item';
 import PdfSheet from '@/components/pdf-drawer';
 import { useClickDrawer } from '@/components/pdf-drawer/hooks';
@@ -102,8 +103,10 @@ function AgentChatBox() {
                   {message.role === MessageType.Assistant &&
                     derivedMessages.length - 1 !== i && (
                       <div>
-                        <div>{message?.data?.tips}</div>
-
+                        <MarkdownContent
+                          content={message?.data?.tips}
+                          loading={false}
+                        ></MarkdownContent>
                         <div>
                           {buildInputList(message)?.map((item) => item.value)}
                         </div>

web/src/pages/agent/constant/pipeline.tsx

@@ -206,6 +206,7 @@ export const initialSplitterValues = {
   chunk_token_size: 512,
   overlapped_percent: 0,
   delimiters: [{ value: '\n' }],
+  image_table_context_window: 0,
 };
 
 export enum Hierarchy {

web/src/pages/agent/debug-content/index.tsx

@@ -1,3 +1,4 @@
+import MarkdownContent from '@/components/next-markdown-content';
 import { ButtonLoading } from '@/components/ui/button';
 import {
   Form,
@@ -234,7 +235,14 @@ const DebugContent = ({
   return (
     <>
       <section>
-        {message?.data?.tips && <div className="mb-2">{message.data.tips}</div>}
+        {message?.data?.tips && (
+          <div className="mb-2">
+            <MarkdownContent
+              content={message?.data?.tips}
+              loading={false}
+            ></MarkdownContent>
+          </div>
+        )}
         <Form {...form}>
           <form onSubmit={form.handleSubmit(onSubmit)} className="space-y-4">
             {parameters.map((x, idx) => {

web/src/pages/agent/form/splitter-form/index.tsx

@@ -22,6 +22,7 @@ const outputList = buildOutputList(initialSplitterValues.outputs);
 
 export const FormSchema = z.object({
   chunk_token_size: z.number(),
+  image_table_context_window: z.number(),
   delimiters: z.array(
     z.object({
       value: z.string().optional(),
@@ -74,6 +75,13 @@ const SplitterForm = ({ node }: INextOperatorForm) => {
           min={0}
           label={t('flow.overlappedPercent')}
         ></SliderInputFormField>
+        <SliderInputFormField
+          name="image_table_context_window"
+          max={256}
+          min={0}
+          label={t('knowledgeConfiguration.imageTableContextWindow')}
+          tooltip={t('knowledgeConfiguration.imageTableContextWindowTip')}
+        ></SliderInputFormField>
         <section>
           <span className="mb-2 inline-block">{t('flow.delimiters')}</span>
           <div className="space-y-4">

web/src/pages/agent/store.ts

@@ -466,7 +466,7 @@ const useGraphStore = create<RFState>()(
         }
       },
       updateSwitchFormData: (source, sourceHandle, target, isConnecting) => {
-        const { updateNodeForm, edges } = get();
+        const { updateNodeForm, edges, getOperatorTypeFromId } = get();
         if (sourceHandle) {
           // A handle will connect to multiple downstream nodes
           let currentHandleTargets = edges
@@ -474,7 +474,8 @@ const useGraphStore = create<RFState>()(
               (x) =>
                 x.source === source &&
                 x.sourceHandle === sourceHandle &&
-                typeof x.target === 'string',
+                typeof x.target === 'string' &&
+                getOperatorTypeFromId(x.target) !== Operator.Placeholder,
             )
             .map((x) => x.target);
 

web/src/pages/agent/utils.ts

@@ -289,10 +289,14 @@ function transformParserParams(params: ParserFormSchemaType) {
 }
 
 function transformSplitterParams(params: SplitterFormSchemaType) {
+  const { image_table_context_window, ...rest } = params;
+  const imageTableContextWindow = Number(image_table_context_window || 0);
   return {
-    ...params,
+    ...rest,
     overlapped_percent: Number(params.overlapped_percent) / 100,
     delimiters: transformObjectArrayToPureArray(params.delimiters, 'value'),
+    table_context_size: imageTableContextWindow,
+    image_context_size: imageTableContextWindow,
 
     // Unset children delimiters if this option is not enabled
     children_delimiters: params.enable_children

web/src/pages/dataset/components/metedata/hooks/use-manage-modal.ts

@@ -1,11 +1,15 @@
 import message from '@/components/ui/message';
 import { useSetModalState } from '@/hooks/common-hooks';
-import { useSetDocumentMeta } from '@/hooks/use-document-request';
+import {
+  DocumentApiAction,
+  useSetDocumentMeta,
+} from '@/hooks/use-document-request';
 import kbService, {
   getMetaDataService,
   updateMetaData,
 } from '@/services/knowledge-service';
-import { useQuery } from '@tanstack/react-query';
+import { useQuery, useQueryClient } from '@tanstack/react-query';
+import { TFunction } from 'i18next';
 import { useCallback, useEffect, useState } from 'react';
 import { useTranslation } from 'react-i18next';
 import { useParams } from 'umi';
@@ -16,12 +20,43 @@ import {
   IMetaDataTableData,
   MetadataOperations,
   ShowManageMetadataModalProps,
-} from './interface';
+} from '../interface';
 export enum MetadataType {
   Manage = 1,
   UpdateSingle = 2,
   Setting = 3,
+  SingleFileSetting = 4,
 }
+
+export const MetadataDeleteMap = (
+  t: TFunction<'translation', undefined>,
+): Record<
+  MetadataType,
+  { title: string; warnFieldText: string; warnValueText: string }
+> => {
+  return {
+    [MetadataType.Manage]: {
+      title: t('common.delete') + ' ' + t('knowledgeDetails.metadata.metadata'),
+      warnFieldText: t('knowledgeDetails.metadata.deleteManageFieldAllWarn'),
+      warnValueText: t('knowledgeDetails.metadata.deleteManageValueAllWarn'),
+    },
+    [MetadataType.Setting]: {
+      title: t('common.delete') + ' ' + t('knowledgeDetails.metadata.metadata'),
+      warnFieldText: t('knowledgeDetails.metadata.deleteSettingFieldWarn'),
+      warnValueText: t('knowledgeDetails.metadata.deleteSettingValueWarn'),
+    },
+    [MetadataType.UpdateSingle]: {
+      title: t('common.delete') + ' ' + t('knowledgeDetails.metadata.metadata'),
+      warnFieldText: t('knowledgeDetails.metadata.deleteManageFieldSingleWarn'),
+      warnValueText: t('knowledgeDetails.metadata.deleteManageValueSingleWarn'),
+    },
+    [MetadataType.SingleFileSetting]: {
+      title: t('common.delete') + ' ' + t('knowledgeDetails.metadata.metadata'),
+      warnFieldText: t('knowledgeDetails.metadata.deleteSettingFieldWarn'),
+      warnValueText: t('knowledgeDetails.metadata.deleteSettingValueWarn'),
+    },
+  };
+};
 export const util = {
   changeToMetaDataTableData(data: IMetaDataReturnType): IMetaDataTableData[] {
     return Object.entries(data).map(([key, value]) => {
@@ -39,10 +74,21 @@ export const util = {
     data: Record<string, string | string[]>,
   ): IMetaDataTableData[] {
     return Object.entries(data).map(([key, value]) => {
+      let thisValue = [] as string[];
+      if (value && Array.isArray(value)) {
+        thisValue = value;
+      } else if (value && typeof value === 'string') {
+        thisValue = [value];
+      } else if (value && typeof value === 'object') {
+        thisValue = [JSON.stringify(value)];
+      } else if (value) {
+        thisValue = [value.toString()];
+      }
+
       return {
         field: key,
         description: '',
-        values: value,
+        values: thisValue,
       } as IMetaDataTableData;
     });
   },
@@ -100,12 +146,42 @@ export const useMetadataOperations = () => {
     }));
   }, []);
 
+  // const addUpdateValue = useCallback(
+  //   (key: string, value: string | string[]) => {
+  //     setOperations((prev) => ({
+  //       ...prev,
+  //       updates: [...prev.updates, { key, value }],
+  //     }));
+  //   },
+  //   [],
+  // );
   const addUpdateValue = useCallback(
-    (key: string, value: string | string[]) => {
-      setOperations((prev) => ({
-        ...prev,
-        updates: [...prev.updates, { key, value }],
-      }));
+    (key: string, originalValue: string, newValue: string) => {
+      setOperations((prev) => {
+        const existsIndex = prev.updates.findIndex(
+          (update) => update.key === key && update.match === originalValue,
+        );
+
+        if (existsIndex > -1) {
+          const updatedUpdates = [...prev.updates];
+          updatedUpdates[existsIndex] = {
+            key,
+            match: originalValue,
+            value: newValue,
+          };
+          return {
+            ...prev,
+            updates: updatedUpdates,
+          };
+        }
+        return {
+          ...prev,
+          updates: [
+            ...prev.updates,
+            { key, match: originalValue, value: newValue },
+          ],
+        };
+      });
     },
     [],
   );
@@ -191,9 +267,14 @@ export const useManageMetaDataModal = (
   const { data, loading } = useFetchMetaDataManageData(type);
 
   const [tableData, setTableData] = useState<IMetaDataTableData[]>(metaData);
-
-  const { operations, addDeleteRow, addDeleteValue, addUpdateValue } =
-    useMetadataOperations();
+  const queryClient = useQueryClient();
+  const {
+    operations,
+    addDeleteRow,
+    addDeleteValue,
+    addUpdateValue,
+    resetOperations,
+  } = useMetadataOperations();
 
   const { setDocumentMeta } = useSetDocumentMeta();
 
@@ -259,11 +340,15 @@ export const useManageMetaDataModal = (
         data: operations,
       });
       if (res.code === 0) {
+        queryClient.invalidateQueries({
+          queryKey: [DocumentApiAction.FetchDocumentList],
+        });
+        resetOperations();
         message.success(t('message.operated'));
         callback();
       }
     },
-    [operations, id, t],
+    [operations, id, t, queryClient, resetOperations],
   );
 
   const handleSaveUpdateSingle = useCallback(
@@ -297,7 +382,26 @@ export const useManageMetaDataModal = (
 
       return data;
     },
-    [tableData, id],
+    [tableData, id, t],
+  );
+
+  const handleSaveSingleFileSettings = useCallback(
+    async (callback: () => void) => {
+      const data = util.tableDataToMetaDataSettingJSON(tableData);
+      if (otherData?.documentId) {
+        const { data: res } = await kbService.documentUpdateMetaData({
+          doc_id: otherData.documentId,
+          metadata: data,
+        });
+        if (res.code === 0) {
+          message.success(t('message.operated'));
+          callback?.();
+        }
+      }
+
+      return data;
+    },
+    [tableData, t, otherData],
   );
 
   const handleSave = useCallback(
@@ -311,12 +415,20 @@ export const useManageMetaDataModal = (
           break;
         case MetadataType.Setting:
           return handleSaveSettings(callback);
+        case MetadataType.SingleFileSetting:
+          return handleSaveSingleFileSettings(callback);
         default:
           handleSaveManage(callback);
           break;
       }
     },
-    [handleSaveManage, type, handleSaveUpdateSingle, handleSaveSettings],
+    [
+      handleSaveManage,
+      type,
+      handleSaveUpdateSingle,
+      handleSaveSettings,
+      handleSaveSingleFileSettings,
+    ],
   );
 
   return {
@@ -371,11 +483,3 @@ export const useManageMetadata = () => {
     config,
   };
 };
-
-export const useManageValues = () => {
-  const [updateValues, setUpdateValues] = useState<{
-    field: string;
-    values: string[];
-  } | null>(null);
-  return { updateValues, setUpdateValues };
-};

web/src/pages/dataset/components/metedata/hooks/use-manage-values-modal.ts

@@ -0,0 +1,208 @@
+import { useCallback, useEffect, useState } from 'react';
+import { useTranslation } from 'react-i18next';
+import { MetadataDeleteMap, MetadataType } from '../hooks/use-manage-modal';
+import { IManageValuesProps, IMetaDataTableData } from '../interface';
+
+export const useManageValues = (props: IManageValuesProps) => {
+  const {
+    data,
+
+    isShowValueSwitch,
+    hideModal,
+    onSave,
+    addUpdateValue,
+    addDeleteValue,
+    existsKeys,
+    type,
+  } = props;
+  const { t } = useTranslation();
+  const [metaData, setMetaData] = useState(data);
+  const [valueError, setValueError] = useState<Record<string, string>>({
+    field: '',
+    values: '',
+  });
+  const [deleteDialogContent, setDeleteDialogContent] = useState({
+    visible: false,
+    title: '',
+    name: '',
+    warnText: '',
+    onOk: () => {},
+    onCancel: () => {},
+  });
+  const hideDeleteModal = () => {
+    setDeleteDialogContent({
+      visible: false,
+      title: '',
+      name: '',
+      warnText: '',
+      onOk: () => {},
+      onCancel: () => {},
+    });
+  };
+
+  // Use functional update to avoid closure issues
+  const handleChange = useCallback(
+    (field: string, value: any) => {
+      if (field === 'field' && existsKeys.includes(value)) {
+        setValueError((prev) => {
+          return {
+            ...prev,
+            field:
+              type === MetadataType.Setting
+                ? t('knowledgeDetails.metadata.fieldExists')
+                : t('knowledgeDetails.metadata.fieldNameExists'),
+          };
+        });
+      } else if (field === 'field' && !existsKeys.includes(value)) {
+        setValueError((prev) => {
+          return {
+            ...prev,
+            field: '',
+          };
+        });
+      }
+      setMetaData((prev) => ({
+        ...prev,
+        [field]: value,
+      }));
+    },
+    [existsKeys, type, t],
+  );
+
+  // Maintain separate state for each input box
+  const [tempValues, setTempValues] = useState<string[]>([...data.values]);
+
+  useEffect(() => {
+    setTempValues([...data.values]);
+    setMetaData(data);
+  }, [data]);
+
+  const handleHideModal = useCallback(() => {
+    hideModal();
+    setMetaData({} as IMetaDataTableData);
+  }, [hideModal]);
+
+  const handleSave = useCallback(() => {
+    if (type === MetadataType.Setting && valueError.field) {
+      return;
+    }
+    if (!metaData.restrictDefinedValues && isShowValueSwitch) {
+      const newMetaData = { ...metaData, values: [] };
+      onSave(newMetaData);
+    } else {
+      onSave(metaData);
+    }
+    handleHideModal();
+  }, [metaData, onSave, handleHideModal, isShowValueSwitch, type, valueError]);
+
+  // Handle value changes, only update temporary state
+  const handleValueChange = useCallback(
+    (index: number, value: string) => {
+      setTempValues((prev) => {
+        if (prev.includes(value)) {
+          setValueError((prev) => {
+            return {
+              ...prev,
+              values: t('knowledgeDetails.metadata.valueExists'),
+            };
+          });
+        } else {
+          setValueError((prev) => {
+            return {
+              ...prev,
+              values: '',
+            };
+          });
+        }
+        const newValues = [...prev];
+        newValues[index] = value;
+
+        return newValues;
+      });
+    },
+    [t],
+  );
+
+  // Handle blur event, synchronize to main state
+  const handleValueBlur = useCallback(() => {
+    // addUpdateValue(metaData.field, [...new Set([...tempValues])]);
+    tempValues.forEach((newValue, index) => {
+      if (index < data.values.length) {
+        const originalValue = data.values[index];
+        if (originalValue !== newValue) {
+          addUpdateValue(metaData.field, originalValue, newValue);
+        }
+      } else {
+        if (newValue) {
+          addUpdateValue(metaData.field, '', newValue);
+        }
+      }
+    });
+    handleChange('values', [...new Set([...tempValues])]);
+  }, [handleChange, tempValues, metaData, data, addUpdateValue]);
+
+  // Handle delete operation
+  const handleDelete = useCallback(
+    (index: number) => {
+      setTempValues((prev) => {
+        const newTempValues = [...prev];
+        addDeleteValue(metaData.field, newTempValues[index]);
+        newTempValues.splice(index, 1);
+        return newTempValues;
+      });
+
+      // Synchronize to main state
+      setMetaData((prev) => {
+        const newMetaDataValues = [...prev.values];
+        newMetaDataValues.splice(index, 1);
+        return {
+          ...prev,
+          values: newMetaDataValues,
+        };
+      });
+    },
+    [addDeleteValue, metaData],
+  );
+
+  const showDeleteModal = (item: string, callback: () => void) => {
+    setDeleteDialogContent({
+      visible: true,
+      title: t('common.delete') + ' ' + t('knowledgeDetails.metadata.value'),
+      name: item,
+      warnText: MetadataDeleteMap(t)[type as MetadataType].warnValueText,
+      onOk: () => {
+        hideDeleteModal();
+        callback();
+      },
+      onCancel: () => {
+        hideDeleteModal();
+      },
+    });
+  };
+
+  // Handle adding new value
+  const handleAddValue = useCallback(() => {
+    setTempValues((prev) => [...new Set([...prev, ''])]);
+
+    // Synchronize to main state
+    setMetaData((prev) => ({
+      ...prev,
+      values: [...new Set([...prev.values, ''])],
+    }));
+  }, []);
+
+  return {
+    metaData,
+    tempValues,
+    valueError,
+    deleteDialogContent,
+    handleChange,
+    handleValueChange,
+    handleValueBlur,
+    handleDelete,
+    handleAddValue,
+    showDeleteModal,
+    handleSave,
+    handleHideModal,
+  };
+};

web/src/pages/dataset/components/metedata/interface.ts

@@ -39,6 +39,7 @@ export type IManageModalProps = {
 
 export interface IManageValuesProps {
   title: ReactNode;
+  existsKeys: string[];
   visible: boolean;
   isEditField?: boolean;
   isAddValue?: boolean;
@@ -46,9 +47,14 @@ export interface IManageValuesProps {
   isShowValueSwitch?: boolean;
   isVerticalShowValue?: boolean;
   data: IMetaDataTableData;
+  type: MetadataType;
   hideModal: () => void;
   onSave: (data: IMetaDataTableData) => void;
-  addUpdateValue: (key: string, value: string | string[]) => void;
+  addUpdateValue: (
+    key: string,
+    originalValue: string,
+    newValue: string,
+  ) => void;
   addDeleteValue: (key: string, value: string) => void;
 }
 
@@ -59,7 +65,8 @@ interface DeleteOperation {
 
 interface UpdateOperation {
   key: string;
-  value: string | string[];
+  match: string;
+  value: string;
 }
 
 export interface MetadataOperations {

web/src/pages/dataset/components/metedata/manage-modal.tsx

@@ -25,11 +25,16 @@ import {
   useReactTable,
 } from '@tanstack/react-table';
 import { Plus, Settings, Trash2 } from 'lucide-react';
-import { useCallback, useMemo, useState } from 'react';
+import { useCallback, useEffect, useMemo, useState } from 'react';
 import { useTranslation } from 'react-i18next';
-import { MetadataType, useManageMetaDataModal } from './hook';
+import {
+  MetadataDeleteMap,
+  MetadataType,
+  useManageMetaDataModal,
+} from './hooks/use-manage-modal';
 import { IManageModalProps, IMetaDataTableData } from './interface';
 import { ManageValuesModal } from './manage-values-modal';
+
 export const ManageMetadataModal = (props: IManageModalProps) => {
   const {
     title,
@@ -54,6 +59,7 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
     values: [],
   });
 
+  const [currentValueIndex, setCurrentValueIndex] = useState<number>(0);
   const [deleteDialogContent, setDeleteDialogContent] = useState({
     visible: false,
     title: '',
@@ -94,12 +100,12 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
       description: '',
       values: [],
     });
-    // setCurrentValueIndex(tableData.length || 0);
+    setCurrentValueIndex(tableData.length || 0);
     showManageValuesModal();
   };
   const handleEditValueRow = useCallback(
-    (data: IMetaDataTableData) => {
-      // setCurrentValueIndex(index);
+    (data: IMetaDataTableData, index: number) => {
+      setCurrentValueIndex(index);
       setValueData(data);
       showManageValuesModal();
     },
@@ -133,7 +139,8 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
           const values = row.getValue('values') as Array<string>;
           return (
             <div className="flex items-center gap-1">
-              {values.length > 0 &&
+              {Array.isArray(values) &&
+                values.length > 0 &&
                 values
                   .filter((value: string, index: number) => index < 2)
                   ?.map((value: string) => {
@@ -153,10 +160,28 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
                               variant={'delete'}
                               className="p-0 bg-transparent"
                               onClick={() => {
-                                handleDeleteSingleValue(
-                                  row.getValue('field'),
-                                  value,
-                                );
+                                setDeleteDialogContent({
+                                  visible: true,
+                                  title:
+                                    t('common.delete') +
+                                    ' ' +
+                                    t('knowledgeDetails.metadata.value'),
+                                  name: value,
+                                  warnText:
+                                    MetadataDeleteMap(t)[
+                                      metadataType as MetadataType
+                                    ].warnValueText,
+                                  onOk: () => {
+                                    hideDeleteModal();
+                                    handleDeleteSingleValue(
+                                      row.getValue('field'),
+                                      value,
+                                    );
+                                  },
+                                  onCancel: () => {
+                                    hideDeleteModal();
+                                  },
+                                });
                               }}
                             >
                               <Trash2 />
@@ -166,7 +191,7 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
                       </Button>
                     );
                   })}
-              {values.length > 2 && (
+              {Array.isArray(values) && values.length > 2 && (
                 <div className="text-text-secondary self-end">...</div>
               )}
             </div>
@@ -185,7 +210,7 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
               variant={'ghost'}
               className="bg-transparent px-1 py-0"
               onClick={() => {
-                handleEditValueRow(row.original);
+                handleEditValueRow(row.original, row.index);
               }}
             >
               <Settings />
@@ -197,11 +222,14 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
                 setDeleteDialogContent({
                   visible: true,
                   title:
-                    t('common.delete') +
-                    ' ' +
-                    t('knowledgeDetails.metadata.metadata'),
+                    // t('common.delete') +
+                    // ' ' +
+                    // t('knowledgeDetails.metadata.metadata')
+                    MetadataDeleteMap(t)[metadataType as MetadataType].title,
                   name: row.getValue('field'),
-                  warnText: t('knowledgeDetails.metadata.deleteWarn'),
+                  warnText:
+                    MetadataDeleteMap(t)[metadataType as MetadataType]
+                      .warnFieldText,
                   onOk: () => {
                     hideDeleteModal();
                     handleDeleteSingleRow(row.getValue('field'));
@@ -240,15 +268,29 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
     getFilteredRowModel: getFilteredRowModel(),
     manualPagination: true,
   });
-
+  const [shouldSave, setShouldSave] = useState(false);
   const handleSaveValues = (data: IMetaDataTableData) => {
     setTableData((prev) => {
-      //If the keys are the same, they need to be merged.
-      const fieldMap = new Map<string, any>();
+      let newData;
+      if (currentValueIndex >= prev.length) {
+        // Add operation
+        newData = [...prev, data];
+      } else {
+        // Edit operation
+        newData = prev.map((item, index) => {
+          if (index === currentValueIndex) {
+            return data;
+          }
+          return item;
+        });
+      }
 
-      prev.forEach((item) => {
+      // Deduplicate by field and merge values
+      const fieldMap = new Map<string, IMetaDataTableData>();
+      newData.forEach((item) => {
         if (fieldMap.has(item.field)) {
-          const existingItem = fieldMap.get(item.field);
+          // Merge values if field exists
+          const existingItem = fieldMap.get(item.field)!;
           const mergedValues = [
             ...new Set([...existingItem.values, ...item.values]),
           ];
@@ -258,20 +300,26 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
         }
       });
 
-      if (fieldMap.has(data.field)) {
-        const existingItem = fieldMap.get(data.field);
-        const mergedValues = [
-          ...new Set([...existingItem.values, ...data.values]),
-        ];
-        fieldMap.set(data.field, { ...existingItem, values: mergedValues });
-      } else {
-        fieldMap.set(data.field, data);
-      }
-
       return Array.from(fieldMap.values());
     });
+    setShouldSave(true);
   };
 
+  useEffect(() => {
+    if (shouldSave) {
+      const timer = setTimeout(() => {
+        handleSave({ callback: () => {} });
+        setShouldSave(false);
+      }, 0);
+
+      return () => clearTimeout(timer);
+    }
+  }, [tableData, shouldSave, handleSave]);
+
+  const existsKeys = useMemo(() => {
+    return tableData.map((item) => item.field);
+  }, [tableData]);
+
   return (
     <>
       <Modal
@@ -352,11 +400,14 @@ export const ManageMetadataModal = (props: IManageModalProps) => {
         <ManageValuesModal
           title={
             <div>
-              {metadataType === MetadataType.Setting
+              {metadataType === MetadataType.Setting ||
+              metadataType === MetadataType.SingleFileSetting
                 ? t('knowledgeDetails.metadata.fieldSetting')
                 : t('knowledgeDetails.metadata.editMetadata')}
             </div>
           }
+          type={metadataType}
+          existsKeys={existsKeys}
           visible={manageValuesVisible}
           hideModal={hideManageValuesModal}
           data={valueData}

web/src/pages/dataset/components/metedata/manage-values-modal.tsx

@@ -1,13 +1,19 @@
+import {
+  ConfirmDeleteDialog,
+  ConfirmDeleteDialogNode,
+} from '@/components/confirm-delete-dialog';
 import EditTag from '@/components/edit-tag';
 import { Button } from '@/components/ui/button';
+import { FormLabel } from '@/components/ui/form';
 import { Input } from '@/components/ui/input';
 import { Modal } from '@/components/ui/modal/modal';
 import { Switch } from '@/components/ui/switch';
 import { Textarea } from '@/components/ui/textarea';
 import { Plus, Trash2 } from 'lucide-react';
-import { memo, useCallback, useEffect, useState } from 'react';
+import { memo } from 'react';
 import { useTranslation } from 'react-i18next';
-import { IManageValuesProps, IMetaDataTableData } from './interface';
+import { useManageValues } from './hooks/use-manage-values-modal';
+import { IManageValuesProps } from './interface';
 
 // Create a separate input component, wrapped with memo to avoid unnecessary re-renders
 const ValueInputItem = memo(
@@ -52,102 +58,29 @@ const ValueInputItem = memo(
 export const ManageValuesModal = (props: IManageValuesProps) => {
   const {
     title,
-    data,
     isEditField,
     visible,
     isAddValue,
     isShowDescription,
     isShowValueSwitch,
     isVerticalShowValue,
-    hideModal,
-    onSave,
-    addUpdateValue,
-    addDeleteValue,
   } = props;
-  const [metaData, setMetaData] = useState(data);
+  const {
+    metaData,
+    tempValues,
+    valueError,
+    deleteDialogContent,
+    handleChange,
+    handleValueChange,
+    handleValueBlur,
+    handleDelete,
+    handleAddValue,
+    showDeleteModal,
+    handleSave,
+    handleHideModal,
+  } = useManageValues(props);
   const { t } = useTranslation();
 
-  // Use functional update to avoid closure issues
-  const handleChange = useCallback((field: string, value: any) => {
-    setMetaData((prev) => ({
-      ...prev,
-      [field]: value,
-    }));
-  }, []);
-
-  // Maintain separate state for each input box
-  const [tempValues, setTempValues] = useState<string[]>([...data.values]);
-
-  useEffect(() => {
-    setTempValues([...data.values]);
-    setMetaData(data);
-  }, [data]);
-
-  const handleHideModal = useCallback(() => {
-    hideModal();
-    setMetaData({} as IMetaDataTableData);
-  }, [hideModal]);
-
-  const handleSave = useCallback(() => {
-    if (!metaData.restrictDefinedValues && isShowValueSwitch) {
-      const newMetaData = { ...metaData, values: [] };
-      onSave(newMetaData);
-    } else {
-      onSave(metaData);
-    }
-    handleHideModal();
-  }, [metaData, onSave, handleHideModal, isShowValueSwitch]);
-
-  // Handle value changes, only update temporary state
-  const handleValueChange = useCallback((index: number, value: string) => {
-    setTempValues((prev) => {
-      const newValues = [...prev];
-      newValues[index] = value;
-
-      return newValues;
-    });
-  }, []);
-
-  // Handle blur event, synchronize to main state
-  const handleValueBlur = useCallback(() => {
-    addUpdateValue(metaData.field, [...new Set([...tempValues])]);
-    handleChange('values', [...new Set([...tempValues])]);
-  }, [handleChange, tempValues, metaData, addUpdateValue]);
-
-  // Handle delete operation
-  const handleDelete = useCallback(
-    (index: number) => {
-      setTempValues((prev) => {
-        const newTempValues = [...prev];
-        addDeleteValue(metaData.field, newTempValues[index]);
-        newTempValues.splice(index, 1);
-        return newTempValues;
-      });
-
-      // Synchronize to main state
-      setMetaData((prev) => {
-        const newMetaDataValues = [...prev.values];
-        newMetaDataValues.splice(index, 1);
-        return {
-          ...prev,
-          values: newMetaDataValues,
-        };
-      });
-    },
-    [addDeleteValue, metaData],
-  );
-
-  // Handle adding new value
-  const handleAddValue = useCallback(() => {
-    setTempValues((prev) => [...new Set([...prev, ''])]);
-
-    // Synchronize to main state
-    setMetaData((prev) => ({
-      ...prev,
-      values: [...new Set([...prev.values, ''])],
-    }));
-  }, []);
-
   return (
     <Modal
       title={title}
@@ -172,15 +105,24 @@ export const ManageValuesModal = (props: IManageValuesProps) => {
               <Input
                 value={metaData.field}
                 onChange={(e) => {
-                  handleChange('field', e.target?.value || '');
+                  const value = e.target?.value || '';
+                  if (/^[a-zA-Z_]*$/.test(value)) {
+                    handleChange('field', value);
+                  }
                 }}
               />
+              <div className="text-state-error text-sm">{valueError.field}</div>
             </div>
           </div>
         )}
         {isShowDescription && (
           <div className="flex flex-col gap-2">
-            <div>{t('knowledgeDetails.metadata.description')}</div>
+            <FormLabel
+              className="text-text-primary text-base"
+              tooltip={t('knowledgeDetails.metadata.descriptionTip')}
+            >
+              {t('knowledgeDetails.metadata.description')}
+            </FormLabel>
             <div>
               <Textarea
                 value={metaData.description}
@@ -193,7 +135,12 @@ export const ManageValuesModal = (props: IManageValuesProps) => {
         )}
         {isShowValueSwitch && (
           <div className="flex flex-col gap-2">
-            <div>{t('knowledgeDetails.metadata.restrictDefinedValues')}</div>
+            <FormLabel
+              className="text-text-primary text-base"
+              tooltip={t('knowledgeDetails.metadata.restrictTDefinedValuesTip')}
+            >
+              {t('knowledgeDetails.metadata.restrictDefinedValues')}
+            </FormLabel>
             <div>
               <Switch
                 checked={metaData.restrictDefinedValues || false}
@@ -230,7 +177,11 @@ export const ManageValuesModal = (props: IManageValuesProps) => {
                       item={item}
                       index={index}
                       onValueChange={handleValueChange}
-                      onDelete={handleDelete}
+                      onDelete={(idx: number) => {
+                        showDeleteModal(item, () => {
+                          handleDelete(idx);
+                        });
+                      }}
                       onBlur={handleValueBlur}
                     />
                   );
@@ -240,11 +191,41 @@ export const ManageValuesModal = (props: IManageValuesProps) => {
             {!isVerticalShowValue && (
               <EditTag
                 value={metaData.values}
-                onChange={(value) => handleChange('values', value)}
+                onChange={(value) => {
+                  // find deleted value
+                  const item = metaData.values.find(
+                    (item) => !value.includes(item),
+                  );
+                  if (item) {
+                    showDeleteModal(item, () => {
+                      // handleDelete(idx);
+                      handleChange('values', value);
+                    });
+                  } else {
+                    handleChange('values', value);
+                  }
+                }}
               />
             )}
+            <div className="text-state-error text-sm">{valueError.values}</div>
           </div>
         )}
+        {deleteDialogContent.visible && (
+          <ConfirmDeleteDialog
+            open={deleteDialogContent.visible}
+            onCancel={deleteDialogContent.onCancel}
+            onOk={deleteDialogContent.onOk}
+            title={deleteDialogContent.title}
+            content={{
+              node: (
+                <ConfirmDeleteDialogNode
+                  name={deleteDialogContent.name}
+                  warnText={deleteDialogContent.warnText}
+                />
+              ),
+            }}
+          />
+        )}
       </div>
     </Modal>
   );

web/src/pages/dataset/dataset-setting/configuration/common-item.tsx

@@ -35,7 +35,8 @@ import {
   MetadataType,
   useManageMetadata,
   util,
-} from '../../components/metedata/hook';
+} from '../../components/metedata/hooks/use-manage-modal';
+import { IMetaDataReturnJSONSettings } from '../../components/metedata/interface';
 import { ManageMetadataModal } from '../../components/metedata/manage-modal';
 import {
   useHandleKbEmbedding,
@@ -359,7 +360,13 @@ export function OverlappedPercent() {
   );
 }
 
-export function AutoMetadata() {
+export function AutoMetadata({
+  type = MetadataType.Setting,
+  otherData,
+}: {
+  type?: MetadataType;
+  otherData?: Record<string, any>;
+}) {
   // get metadata field
   const form = useFormContext();
   const {
@@ -369,6 +376,7 @@ export function AutoMetadata() {
     tableData,
     config: metadataConfig,
   } = useManageMetadata();
+
   const autoMetadataField: FormFieldConfig = {
     name: 'parser_config.enable_metadata',
     label: t('knowledgeConfiguration.autoMetadata'),
@@ -379,6 +387,7 @@ export function AutoMetadata() {
     render: (fieldProps: ControllerRenderProps) => (
       <div className="flex items-center justify-between">
         <Button
+          type="button"
           variant="ghost"
           onClick={() => {
             const metadata = form.getValues('parser_config.metadata');
@@ -387,7 +396,8 @@ export function AutoMetadata() {
             showManageMetadataModal({
               metadata: tableMetaData,
               isCanAdd: true,
-              type: MetadataType.Setting,
+              type: type,
+              record: otherData,
             });
           }}
         >
@@ -403,6 +413,10 @@ export function AutoMetadata() {
       </div>
     ),
   };
+
+  const handleSaveMetadata = (data?: IMetaDataReturnJSONSettings) => {
+    form.setValue('parser_config.metadata', data || []);
+  };
   return (
     <>
       <RenderField field={autoMetadataField} />
@@ -431,8 +445,8 @@ export function AutoMetadata() {
           isShowDescription={true}
           isShowValueSwitch={true}
           isVerticalShowValue={false}
-          success={(data) => {
-            form.setValue('parser_config.metadata', data || []);
+          success={(data?: IMetaDataReturnJSONSettings) => {
+            handleSaveMetadata(data);
           }}
         />
       )}

web/src/pages/dataset/dataset-setting/form-schema.ts

@@ -96,7 +96,7 @@ export const formSchema = z
           )
           .optional(),
         enable_metadata: z.boolean().optional(),
-        llm_id: z.string().optional(),
+        llm_id: z.string().min(1, { message: 'Indexing model is required' }),
       })
       .optional(),
     pagerank: z.number(),

web/src/pages/dataset/dataset/index.tsx

@@ -16,7 +16,10 @@ import { useFetchKnowledgeBaseConfiguration } from '@/hooks/use-knowledge-reques
 import { Pen, Upload } from 'lucide-react';
 import { useMemo } from 'react';
 import { useTranslation } from 'react-i18next';
-import { MetadataType, useManageMetadata } from '../components/metedata/hook';
+import {
+  MetadataType,
+  useManageMetadata,
+} from '../components/metedata/hooks/use-manage-modal';
 import { ManageMetadataModal } from '../components/metedata/manage-modal';
 import { DatasetTable } from './dataset-table';
 import Generate from './generate-button/generate';

web/src/pages/dataset/dataset/use-dataset-table-columns.tsx

@@ -16,7 +16,10 @@ import { formatDate } from '@/utils/date';
 import { ColumnDef } from '@tanstack/table-core';
 import { ArrowUpDown, MonitorUp } from 'lucide-react';
 import { useTranslation } from 'react-i18next';
-import { MetadataType, util } from '../components/metedata/hook';
+import {
+  MetadataType,
+  util,
+} from '../components/metedata/hooks/use-manage-modal';
 import { ShowManageMetadataModalProps } from '../components/metedata/interface';
 import { DatasetActionCell } from './dataset-action-cell';
 import { ParsingStatusCell } from './parsing-status-cell';

web/src/pages/dataset/dataset/use-select-filters.ts

@@ -1,8 +1,13 @@
-import { FilterCollection } from '@/components/list-filter-bar/interface';
+import {
+  FilterCollection,
+  FilterType,
+} from '@/components/list-filter-bar/interface';
 import { useTranslate } from '@/hooks/common-hooks';
 import { useGetDocumentFilter } from '@/hooks/use-document-request';
 import { useMemo } from 'react';
 
+export const EMPTY_METADATA_FIELD = 'empty_metadata';
+
 export function useSelectDatasetFilters() {
   const { t } = useTranslate('knowledgeDetails');
   const { filter, onOpenChange } = useGetDocumentFilter();
@@ -17,34 +22,52 @@ export function useSelectDatasetFilters() {
     }
   }, [filter.suffix]);
   const fileStatus = useMemo(() => {
+    let list = [] as FilterType[];
     if (filter.run_status) {
-      return Object.keys(filter.run_status).map((x) => ({
+      list = Object.keys(filter.run_status).map((x) => ({
         id: x,
         label: t(`runningStatus${x}`),
         count: filter.run_status[x as unknown as number],
       }));
     }
-  }, [filter.run_status, t]);
+    if (filter.metadata) {
+      const emptyMetadata = filter.metadata?.empty_metadata;
+      if (emptyMetadata) {
+        list.push({
+          id: EMPTY_METADATA_FIELD,
+          label: t('emptyMetadata'),
+          count: emptyMetadata.true,
+        });
+      }
+    }
+    return list;
+  }, [filter.run_status, filter.metadata, t]);
   const metaDataList = useMemo(() => {
     if (filter.metadata) {
-      return Object.keys(filter.metadata).map((x) => ({
-        id: x.toString(),
-        field: x.toString(),
-        label: x.toString(),
-        list: Object.keys(filter.metadata[x]).map((y) => ({
-          id: y.toString(),
-          field: y.toString(),
-          label: y.toString(),
-          value: [y],
-          count: filter.metadata[x][y],
-        })),
-        count: Object.keys(filter.metadata[x]).reduce(
-          (acc, cur) => acc + filter.metadata[x][cur],
-          0,
-        ),
-      }));
+      const list = Object.keys(filter.metadata)
+        ?.filter((m) => m !== EMPTY_METADATA_FIELD)
+        ?.map((x) => {
+          return {
+            id: x.toString(),
+            field: x.toString(),
+            label: x.toString(),
+            list: Object.keys(filter.metadata[x]).map((y) => ({
+              id: y.toString(),
+              field: y.toString(),
+              label: y.toString(),
+              value: [y],
+              count: filter.metadata[x][y],
+            })),
+            count: Object.keys(filter.metadata[x]).reduce(
+              (acc, cur) => acc + filter.metadata[x][cur],
+              0,
+            ),
+          };
+        });
+      return list;
     }
   }, [filter.metadata]);
+
   const filters: FilterCollection[] = useMemo(() => {
     return [
       { field: 'type', label: 'File Type', list: fileTypes },

web/src/pages/dataset/process-log-modal.tsx

@@ -1,6 +1,11 @@
 import FileStatusBadge from '@/components/file-status-badge';
 import { Button } from '@/components/ui/button';
 import { Modal } from '@/components/ui/modal/modal';
+import {
+  Tooltip,
+  TooltipContent,
+  TooltipTrigger,
+} from '@/components/ui/tooltip';
 import { RunningStatusMap } from '@/constants/knowledge';
 import { useTranslate } from '@/hooks/common-hooks';
 import React, { useMemo } from 'react';
@@ -40,7 +45,14 @@ const InfoItem: React.FC<{
   return (
     <div className={`flex flex-col mb-4 ${className}`}>
       <span className="text-text-secondary text-sm">{label}</span>
-      <span className="text-text-primary mt-1">{value}</span>
+      <Tooltip>
+        <TooltipTrigger asChild>
+          <span className="text-text-primary mt-1 truncate max-w-[200px]">
+            {value}
+          </span>
+        </TooltipTrigger>
+        <TooltipContent>{value}</TooltipContent>
+      </Tooltip>
     </div>
   );
 };
@@ -70,9 +82,7 @@ const ProcessLogModal: React.FC<ProcessLogModalProps> = ({
 }) => {
   const { t } = useTranslate('knowledgeDetails');
   const blackKeyList = [''];
-  console.log('logInfo', initData);
   const logInfo = useMemo(() => {
-    console.log('logInfo', initData);
     return initData;
   }, [initData]);
 

web/src/pages/datasets/dataset-creating-dialog.tsx

@@ -17,6 +17,7 @@ import {
 } from '@/components/ui/form';
 import { Input } from '@/components/ui/input';
 import { FormLayout } from '@/constants/form';
+import { useFetchTenantInfo } from '@/hooks/use-user-setting-request';
 import { IModalProps } from '@/interfaces/common';
 import { zodResolver } from '@hookform/resolvers/zod';
 import { useEffect } from 'react';
@@ -33,6 +34,7 @@ const FormId = 'dataset-creating-form';
 
 export function InputForm({ onOk }: IModalProps<any>) {
   const { t } = useTranslation();
+  const { data: tenantInfo } = useFetchTenantInfo();
 
   const FormSchema = z
     .object({
@@ -80,7 +82,7 @@ export function InputForm({ onOk }: IModalProps<any>) {
       name: '',
       parseType: 1,
       parser_id: '',
-      embd_id: '',
+      embd_id: tenantInfo?.embd_id,
     },
   });
 

web/src/services/knowledge-service.ts

@@ -48,10 +48,10 @@ const {
   traceRaptor,
   check_embedding,
   kbUpdateMetaData,
+  documentUpdateMetaData,
 } = api;
 
 const methods = {
-  // 知识库管理
   createKb: {
     url: create_kb,
     method: 'post',
@@ -220,6 +220,10 @@ const methods = {
     url: kbUpdateMetaData,
     method: 'post',
   },
+  documentUpdateMetaData: {
+    url: documentUpdateMetaData,
+    method: 'post',
+  },
   // getMetaData: {
   //   url: getMetaData,
   //   method: 'get',
@@ -263,7 +267,7 @@ export const documentFilter = (kb_id: string) =>
 export const getMetaDataService = ({ kb_id }: { kb_id: string }) =>
   request.post(api.getMetaData, { data: { kb_id } });
 export const updateMetaData = ({ kb_id, data }: { kb_id: string; data: any }) =>
-  request.post(api.updateMetaData, { data: { kb_id, data } });
+  request.post(api.updateMetaData, { data: { kb_id, ...data } });
 
 export const listDataPipelineLogDocument = (
   params?: IFetchKnowledgeListRequestParams,

web/src/utils/api.ts

@@ -80,6 +80,7 @@ export default {
   getMetaData: `${api_host}/document/metadata/summary`,
   updateMetaData: `${api_host}/document/metadata/update`,
   kbUpdateMetaData: `${api_host}/kb/update_metadata_setting`,
+  documentUpdateMetaData: `${api_host}/document/update_metadata_setting`,
 
   // tags
   listTag: (knowledgeId: string) => `${api_host}/kb/${knowledgeId}/tags`,